To see the other types of publications on this topic, follow the link: Content-based information retrieval.
Dissertations / Theses on the topic 'Content-based information retrieval'
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 50 dissertations / theses for your research on the topic 'Content-based information retrieval.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Browse dissertations / theses on a wide variety of disciplines and organise your bibliography correctly.
1
Chafik, Sanaa. "Machine learning techniques for content-based information retrieval." Thesis, Université Paris-Saclay (ComUE), 2017. http://www.theses.fr/2017SACLL008/document.
Full textAbstract:
Avec l’évolution des technologies numériques et la prolifération d'internet, la quantité d’information numérique a considérablement évolué. La recherche par similarité (ou recherche des plus proches voisins) est une problématique que plusieurs communautés de recherche ont tenté de résoudre. Les systèmes de recherche par le contenu de l’information constituent l’une des solutions prometteuses à ce problème. Ces systèmes sont composés essentiellement de trois unités fondamentales, une unité de représentation des données pour l’extraction des primitives, une unité d’indexation multidimensionnelle pour la structuration de l’espace des primitives, et une unité de recherche des plus proches voisins pour la recherche des informations similaires. L’information (image, texte, audio, vidéo) peut être représentée par un vecteur multidimensionnel décrivant le contenu global des données d’entrée. La deuxième unité consiste à structurer l’espace des primitives dans une structure d’index, où la troisième unité -la recherche par similarité- est effective.Dans nos travaux de recherche, nous proposons trois systèmes de recherche par le contenu de plus proches voisins. Les trois approches sont non supervisées, et donc adaptées aux données étiquetées et non étiquetées. Elles sont basées sur le concept du hachage pour une recherche efficace multidimensionnelle des plus proches voisins. Contrairement aux approches de hachage existantes, qui sont binaires, les approches proposées fournissent des structures d’index avec un hachage réel. Bien que les approches de hachage binaires fournissent un bon compromis qualité-temps de calcul, leurs performances en termes de qualité (précision) se dégradent en raison de la perte d’information lors du processus de binarisation. À l'opposé, les approches de hachage réel fournissent une bonne qualité de recherche avec une meilleure approximation de l’espace d’origine, mais induisent en général un surcoût en temps de calcul.Ce dernier problème est abordé dans la troisième contribution. Les approches proposées sont classifiées en deux catégories, superficielle et profonde. Dans la première catégorie, on propose deux techniques de hachage superficiel, intitulées Symmetries of the Cube Locality sensitive hashing (SC-LSH) et Cluster-Based Data Oriented Hashing (CDOH), fondées respectivement sur le hachage aléatoire et l’apprentissage statistique superficiel. SCLSH propose une solution au problème de l’espace mémoire rencontré par la plupart des approches de hachage aléatoire, en considérant un hachage semi-aléatoire réduisant partiellement l’effet aléatoire, et donc l’espace mémoire, de ces dernières, tout en préservant leur efficacité pour la structuration des espaces hétérogènes. La seconde technique, CDOH, propose d’éliminer l’effet aléatoire en combinant des techniques d’apprentissage non-supervisé avec le concept de hachage. CDOH fournit de meilleures performances en temps de calcul, en espace mémoire et en qualité de recherche.La troisième contribution est une approche de hachage basée sur les réseaux de neurones profonds appelée "Unsupervised Deep Neuron-per-Neuron Hashing" (UDN2H). UDN2H propose une indexation individuelle de la sortie de chaque neurone de la couche centrale d’un modèle non supervisé. Ce dernier est un auto-encodeur profond capturant une structure individuelle de haut niveau de chaque neurone de sortie.Nos trois approches, SC-LSH, CDOH et UDN2H, ont été proposées séquentiellement durant cette thèse, avec un niveau croissant, en termes de la complexité des modèles développés, et en termes de la qualité de recherche obtenue sur de grandes bases de données d'informationThe amount of media data is growing at high speed with the fast growth of Internet and media resources. Performing an efficient similarity (nearest neighbor) search in such a large collection of data is a very challenging problem that the scientific community has been attempting to tackle. One of the most promising solutions to this fundamental problem is Content-Based Media Retrieval (CBMR) systems. The latter are search systems that perform the retrieval task in large media databases based on the content of the data. CBMR systems consist essentially of three major units, a Data Representation unit for feature representation learning, a Multidimensional Indexing unit for structuring the resulting feature space, and a Nearest Neighbor Search unit to perform efficient search. Media data (i.e. image, text, audio, video, etc.) can be represented by meaningful numeric information (i.e. multidimensional vector), called Feature Description, describing the overall content of the input data. The task of the second unit is to structure the resulting feature descriptor space into an index structure, where the third unit, effective nearest neighbor search, is performed.In this work, we address the problem of nearest neighbor search by proposing three Content-Based Media Retrieval approaches. Our three approaches are unsupervised, and thus can adapt to both labeled and unlabeled real-world datasets. They are based on a hashing indexing scheme to perform effective high dimensional nearest neighbor search. Unlike most recent existing hashing approaches, which favor indexing in Hamming space, our proposed methods provide index structures adapted to a real-space mapping. Although Hamming-based hashing methods achieve good accuracy-speed tradeoff, their accuracy drops owing to information loss during the binarization process. By contrast, real-space hashing approaches provide a more accurate approximation in the mapped real-space as they avoid the hard binary approximations.Our proposed approaches can be classified into shallow and deep approaches. In the former category, we propose two shallow hashing-based approaches namely, "Symmetries of the Cube Locality Sensitive Hashing" (SC-LSH) and "Cluster-based Data Oriented Hashing" (CDOH), based respectively on randomized-hashing and shallow learning-to-hash schemes. The SC-LSH method provides a solution to the space storage problem faced by most randomized-based hashing approaches. It consists of a semi-random scheme reducing partially the randomness effect of randomized hashing approaches, and thus the memory storage problem, while maintaining their efficiency in structuring heterogeneous spaces. The CDOH approach proposes to eliminate the randomness effect by combining machine learning techniques with the hashing concept. The CDOH outperforms the randomized hashing approaches in terms of computation time, memory space and search accuracy.The third approach is a deep learning-based hashing scheme, named "Unsupervised Deep Neuron-per-Neuron Hashing" (UDN2H). The UDN2H approach proposes to index individually the output of each neuron of the top layer of a deep unsupervised model, namely a Deep Autoencoder, with the aim of capturing the high level individual structure of each neuron output.Our three approaches, SC-LSH, CDOH and UDN2H, were proposed sequentially as the thesis was progressing, with an increasing level of complexity in terms of the developed models, and in terms of the effectiveness and the performances obtained on large real-world datasets
2
Ren, Feng Hui. "Multi-image query content-based image retrieval." Access electronically, 2006. http://www.library.uow.edu.au/adt-NWU/public/adt-NWU20070103.143624/index.html.
Full text
3
Wang, Lei. "Content based video retrieval via spatial-temporal information discovery." Thesis, Robert Gordon University, 2013. http://hdl.handle.net/10059/1119.
Full textAbstract:
Content based video retrieval (CBVR) has been strongly motivated by a variety of realworld applications. Most state-of-the-art CBVR systems are built based on Bag-of-visual- Words (BovW) framework for visual resources representation and access. The framework, however, ignores spatial and temporal information contained in videos, which plays a fundamental role in unveiling semantic meanings. The information includes not only the spatial layout of visual content on a still frame (image), but also temporal changes across the sequential frames. Specially, spatially and temporally co-occurring visual words, which are extracted under the BovW framework, often tend to collaboratively represent objects, scenes, or events in the videos. The spatial and temporal information discovery would be useful to advance the CBVR technology. In this thesis, we propose to explore and analyse the spatial and temporal information from a new perspective: i) co-occurrence of the visual words is formulated as a correlation matrix, ii) spatial proximity and temporal coherence are analytically and empirically studied to re ne this correlation. Following this, a quantitative spatial and temporal correlation (STC) model is de ned. The STC discovered from either the query example (denoted by QC) or the data collection (denoted by DC) are assumed to determine speci- city of the visual words in the retrieval model, i:e: selected Words-Of-Interest are found more important for certain topics. Based on this hypothesis, we utilized the STC matrix to establish a novel visual content similarity measurement method and a query reformulation scheme for the retrieval model. Additionally, the STC also characterizes the context of the visual words, and accordingly a STC-Based context similarity measurement is proposed to detect the synonymous visual words. The method partially solves an inherent error of visual vocabulary under the BovW framework. Systematic experimental evaluations on public TRECVID and CC WEB VIDEO video collections demonstrate that the proposed methods based on the STC can substantially improve retrieval e ectiveness of the BovW framework. The retrieval model based on STC outperforms state-of-the-art CBVR methods on the data collections without storage and computational expense. Furthermore, the rebuilt visual vocabulary in this thesis is more compact and e ective. Above methods can be incorporated together for e ective and e cient CBVR system implementation. Based on the experimental results, it is concluded that the spatial-temporal correlation e ectively approximates the semantical correlation. This discovered correlation approximation can be utilized for both visual content representation and similarity measurement, which are key issues for CBVR technology development.
4
Hemgren, Dan. "Fuzzy Content-Based Audio Retrieval Using Visualization Tools." Thesis, KTH, Skolan för elektroteknik och datavetenskap (EECS), 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-264514.
Full textAbstract:
Music composition and sound design in the digital domain often involves sifting through large collections of audio files to find the right sample. Traditionally, this involves searching through metadata such as filenames and descriptors either via text search or by manually searching through folders. This paper presents a fast, scalable method for implementing a search engine in which the contents of audio files are used as queries to retrieve similar audio files. The presented approach applies visualization tools to speed up retrieval time compared to a simple KD-Tree algorithm. Qualitative and quantitative results are presented and benefits and drawbacks of the approach are discussed. While the qualitative results show promise, they are deemed inconclusive. Via the quantitative results, it is found that the application of UMAP yield an order-of-magnitude speed-up at a loss of accuracy and that the approach scales well with larger datasets.Digital ljuddesign och musikkomposition innebär ofta ett sökande genom stora samlingar av ljudfiler efter rätt sampling. Traditionellt sett innebär detta antingen textsökning via metadata såsom filnamn och tags eller manuell sökning genom filstrukturer. Denna rapport presenterar en snabb, skalbar lösning i form av en sökmotor som möjliggör användandet av en ljudfil för innehållsbaserad sökning som hittar liknande ljudfiler. Den presenterade lösningen använder visualiseringsverktyg för att snabba upp hämtningstiden jämför med enkla KD-tree-algoritmer. Kvalitativa och kvantitativa resultat presenteras och för- och nackdelar med lösningen diskuteras. De kvalitativa resultaten visar på potential men bedöms vara ofullständiga. De kvantitativa resultaten påvisar storleksordningar kortare hämtningstid då UMAP används, dock med sänkt noggrannhet som följd, och lösningen visar sig skala väl med större mängder data.
5
Osodo, Jennifer Akinyi. "An extended vector-based information retrieval system to retrieve e-learning content based on learner models." Thesis, University of Sunderland, 2011. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.542053.
Full text
6
Wang, Ben. "Efficient indexing structures for similarity search in content-based information retrieval." Thesis, University of Essex, 2007. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.438150.
Full text
7
Suyoto, Iman S. H., and ishs@ishs net. "Cross-Domain Content-Based Retrieval of Audio Music through Transcription." RMIT University. Computer Science and Information Technology, 2009. http://adt.lib.rmit.edu.au/adt/public/adt-VIT20090527.092841.
Full textAbstract:
Research in the field of music information retrieval (MIR) is concerned with methods to effectively retrieve a piece of music based on a user's query. An important goal in MIR research is the ability to successfully retrieve music stored as recorded audio using note-based queries. In this work, we consider the searching of musical audio using symbolic queries. We first examined the effectiveness of using a relative pitch approach to represent queries and pieces. Our experimental results revealed that this technique, while effective, is optimal when the whole tune is used as a query. We then suggested an algorithm involving the use of pitch classes in conjunction with the longest common subsequence algorithm between a query and target, also using the whole tune as a query. We also proposed an algorithm that works effectively when only a small part of a tune is used as a query. The algorithm makes use of a sliding window in addition to pitch classes and the longest common subsequence algorithm between a query and target. We examined the algorithm using queries based on the beginning, middle, and ending parts of pieces. We performed experiments on an audio collection and manually-constructed symbolic queries. Our experimental evaluation revealed that our techniques are highly effective, with most queries used in our experiments being able to retrieve a correct answer in the first rank position. In addition, we examined the effectiveness of duration-based features for improving retrieval effectiveness over the use of pitch only. We investigated note durations and inter-onset intervals. For this purpose, we used solely symbolic music so that we could focus on the core of the problem. A relative pitch approach alongside a relative duration representation were used in our experiments. Our experimental results showed that durations fail to significantly improve retrieval effectiveness, whereas inter-onset intervals significantly improve retrieval effectiveness.
8
Cloete, Candice Lynn. "MIRMaid : an interface for a content based Music Information Retrieval test-bed." Thesis, University of Cape Town, 2006. http://pubs.cs.uct.ac.za/archive/00000460/.
Full textAbstract:
Music Information Retrieval (MlR) is the interdisciplinary science of retrieving information from music and includes influences from different areas, like music perception and cognition, music analysis, signal processing, music indexing and information retrieval [Futrelle & Downie, 2003].
To produce the most efficient MlR systems, test-beds are commonly used to test different combinations of parameters against each other. The purpose of this dissertation was to investigate the composition of algorithms for MlR systems by constructing an interface that could form part of a test-bed. It differs from other interfaces and frameworks that are used in MlR test-beds because it is focused on small scale test-beds.
MIRMaid is an acronym for Music Information Retrieval Modular aid and is an interface that allows different content based retrieval tasks to be compared against each other to find optimal combinations of retrieval parameters for specialised problem domains.
The dissertation describes the process of how the MIRMaid interface was developed, modified and refined.
A big challenge was to design the user experiments in a way that considered potential users of the interface while using the test subjects I had at my disposal. I decided to use the simplest queries to highlight basic similarities between novice and potential expert users. The performance of the interface was judged by user ratings on a questionnaire. The interface performed reasonably well with expert users and novice users. Despite these results there were a few interesting observations that were returned from the user experiments related to the experiment design and the task explanations.
Some suggestions are also provided for extending the interface to allow it to be used with other types of data. The possibility is also investigated for using the interface as a tool for simplifying the process of integrating modules from different sources.
9
Ma, Qiang. "Query-free information retrieval based on spatio-temporal criteria and content complementation." 京都大学 (Kyoto University), 2004. http://hdl.handle.net/2433/64945.
Full text
10
Kitahara, Tetsuro. "Computational musical instrument recognition and its application to content-based music information retrieval." 京都大学 (Kyoto University), 2007. http://hdl.handle.net/2433/135955.
Full text
11
Kidambi, Phani Nandan. "A HUMAN-COMPUTER INTEGRATED APPROACH TOWARDS CONTENT BASED IMAGE RETRIEVAL." Wright State University / OhioLINK, 2010. http://rave.ohiolink.edu/etdc/view?acc_num=wright1292647701.
Full text
12
Govindarajan, Hariprasath. "Self-Supervised Representation Learning for Content Based Image Retrieval." Thesis, Linköpings universitet, Statistik och maskininlärning, 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-166223.
Full textAbstract:
Automotive technologies and fully autonomous driving have seen a tremendous growth in recent times and have benefitted from extensive deep learning research. State-of-the-art deep learning methods are largely supervised and require labelled data for training. However, the annotation process for image data is time-consuming and costly in terms of human efforts. It is of interest to find informative samples for labelling by Content Based Image Retrieval (CBIR). Generally, a CBIR method takes a query image as input and returns a set of images that are semantically similar to the query image. The image retrieval is achieved by transforming images to feature representations in a latent space, where it is possible to reason about image similarity in terms of image content. In this thesis, a self-supervised method is developed to learn feature representations of road scenes images. The self-supervised method learns feature representations for images by adapting intermediate convolutional features from an existing deep Convolutional Neural Network (CNN). A contrastive approach based on Noise Contrastive Estimation (NCE) is used to train the feature learning model. For complex images like road scenes where mutiple image aspects can occur simultaneously, it is important to embed all the salient image aspects in the feature representation. To achieve this, the output feature representation is obtained as an ensemble of feature embeddings which are learned by focusing on different image aspects. An attention mechanism is incorporated to encourage each ensemble member to focus on different image aspects. For comparison, a self-supervised model without attention is considered and a simple dimensionality reduction approach using SVD is treated as the baseline. The methods are evaluated on nine different evaluation datasets using CBIR performance metrics. The datasets correspond to different image aspects and concern the images at different spatial levels - global, semi-global and local. The feature representations learned by self-supervised methods are shown to perform better than the SVD approach. Taking into account that no labelled data is required for training, learning representations for road scenes images using self-supervised methods appear to be a promising direction. Usage of multiple query images to emphasize a query intention is investigated and a clear improvement in CBIR performance is observed. It is inconclusive whether the addition of an attentive mechanism impacts CBIR performance. The attention method shows some positive signs based on qualitative analysis and also performs better than other methods for one of the evaluation datasets containing a local aspect. This method for learning feature representations is promising but requires further research involving more diverse and complex image aspects.
13
Tang, Siu-shing. "Integrating distance function learning and support vector machine for content-based image retrieval /." View abstract or full-text, 2006. http://library.ust.hk/cgi/db/thesis.pl?CSED%202006%20TANG.
Full text
14
Wong, Chan Fong. "Content-based image retrieval using color quantization, rectangular segmentation, and relevance feedback." Thesis, University of Macau, 2008. http://umaclib3.umac.mo/record=b1780398.
Full text
15
Chi, Pin-Hao. "Efficient protein tertiary structure retrievals and classifications using content based comparison algorithms." Diss., Columbia, Mo. : University of Missouri-Columbia, 2007. http://hdl.handle.net/10355/4817.
Full textAbstract:
Thesis (Ph. D.)--University of Missouri-Columbia, 2007.The entire dissertation/thesis text is included in the research.pdf file; the official abstract appears in the short.pdf file (which also appears in the research.pdf); a non-technical general description, or public abstract, appears in the public.pdf file. Title from title screen of research.pdf file (viewed on September 19, 2007) Vita. Includes bibliographical references.
16
Vemulapalli, Smita. "Audio-video based handwritten mathematical content recognition." Diss., Georgia Institute of Technology, 2012. http://hdl.handle.net/1853/45958.
Full textAbstract:
Recognizing handwritten mathematical content is a challenging problem, and more so when such content appears in classroom videos. However, given the fact that in such videos the handwritten text and the accompanying audio refer to the same content, a combination of video and audio based recognizer has the potential to significantly improve the content recognition accuracy. This dissertation, using a combination of video and audio based recognizers, focuses on improving the recognition accuracy associated with handwritten mathematical content in such videos.
Our approach makes use of a video recognizer as the primary recognizer and a multi-stage assembly, developed as part of this research, is used to facilitate effective combination with an audio recognizer. Specifically, we address the following challenges related to audio-video based handwritten mathematical content recognition: (1) Video Preprocessing - generates a timestamped sequence of segmented characters from the classroom video in the face of occlusions and shadows caused by the instructor, (2) Ambiguity Detection - determines the subset of input characters that may have been incorrectly recognized by the video based recognizer and forwards this subset for disambiguation, (3) A/V Synchronization - establishes correspondence between the handwritten character and the spoken content, (4) A/V Combination - combines the synchronized outputs from the video and audio based recognizers and generates the final recognized character, and (5) Grammar Assisted A/V Based Mathematical Content Recognition - utilizes a base mathematical speech grammar for both character and structure disambiguation. Experiments conducted using videos recorded in a classroom-like environment demonstrate the significant improvements in recognition accuracy that can be achieved using our techniques.
17
Yapp, Lawrence. "Content-based indexing of MPEG video through the analysis of the accompanying audio /." Thesis, Connect to this title online; UW restricted, 1997. http://hdl.handle.net/1773/5835.
Full text
18
Viet, Tran Linh. "Efficient Image Retrieval with Statistical Color Descriptors." Doctoral thesis, Linköpings universitet, Institutionen för teknik och naturvetenskap, 2003. http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-5002.
Full textAbstract:
Color has been widely used in content-based image retrieval (CBIR) applications. In such applications the color properties of an image are usually characterized by the probability distribution of the colors in the image. A distance measure is then used to measure the (dis-)similarity between images based on the descriptions of their color distributions in order to quickly find relevant images. The development and investigation of statistical methods for robust representations of such distributions, the construction of distance measures between them and their applications in efficient retrieval, browsing, and structuring of very large image databases are the main contributions of the thesis. In particular we have addressed the following problems in CBIR. Firstly, different non-parametric density estimators are used to describe color information for CBIR applications. Kernel-based methods using nonorthogonal bases together with a Gram-Schmidt procedure and the application of the Fourier transform are introduced and compared to previously used histogram-based methods. Our experiments show that efficient use of kernel density estimators improves the retrieval performance of CBIR. The practical problem of how to choose an optimal smoothing parameter for such density estimators as well as the selection of the histogram bin-width for CBIR applications are also discussed. Distance measures between color distributions are then described in a differential geometry-based framework. This allows the incorporation of geometrical features of the underlying color space into the distance measure between the probability distributions. The general framework is illustrated with two examples: Normal distributions and linear representations of distributions. The linear representation of color distributions is then used to derive new compact descriptors for color-based image retrieval. These descriptors are based on the combination of two ideas: Incorporating information from the structure of the color space with information from images and application of projection methods in the space of color distribution and the space of differences between neighboring color distributions. In our experiments we used several image databases containing more than 1,300,000 images. The experiments show that the method developed in this thesis is very fast and that the retrieval performance chievedcompares favorably with existing methods. A CBIR system has been developed and is currently available at http://www.media.itn.liu.se/cse. We also describe color invariant descriptors that can be used to retrieve images of objects independent of geometrical factors and the illumination conditions under which these images were taken. Both statistics- and physics-based methods are proposed and examined. We investigated the interaction between light and material using different physical models and applied the theory of transformation groups to derive geometry color invariants. Using the proposed framework, we are able to construct all independent invariants for a given physical model. The dichromatic reflection model and the Kubelka-Munk model are used as examples for the framework. The proposed color invariant descriptors are then applied to both CBIR, color image segmentation, and color correction applications. In the last chapter of the thesis we describe an industrial application where different color correction methods are used to optimize the layout of a newspaper page.A search engine based, on the methodes discribed in this thesis, can be found at http://pub.ep.liu.se/cse/db/?. Note that the question mark must be included in the address.
19
Meng, Zhao. "A Study on Web Search based on Coordinate Relationships." 京都大学 (Kyoto University), 2016. http://hdl.handle.net/2433/217205.
Full text
20
Gouws, Stephan. "Evaluation and development of conceptual document similarity metrics with content-based recommender applications." Thesis, Stellenbosch : University of Stellenbosch, 2010. http://hdl.handle.net/10019.1/5363.
Full textAbstract:
Thesis (MScEng (Electrical and Electronic Engineering))--University of Stellenbosch, 2010.ENGLISH ABSTRACT: The World Wide Web brought with it an unprecedented level of information overload. Computers are very effective at processing and clustering numerical and binary data, however, the automated conceptual clustering of natural-language data is considerably harder to automate. Most past techniques rely on simple keyword-matching techniques or probabilistic methods to measure semantic relatedness. However, these approaches do not always accurately capture conceptual relatedness as measured by humans. In this thesis we propose and evaluate the use of novel Spreading Activation (SA) techniques for computing semantic relatedness, by modelling the article hyperlink structure of Wikipedia as an associative network structure for knowledge representation. The SA technique is adapted and several problems are addressed for it to function over the Wikipedia hyperlink structure. Inter-concept and inter-document similarity metrics are developed which make use of SA to compute the conceptual similarity between two concepts and between two natural-language documents. We evaluate these approaches over two document similarity datasets and achieve results which compare favourably with the state of the art. Furthermore, document preprocessing techniques are evaluated in terms of the performance gain these techniques can have on the well-known cosine document similarity metric and the Normalised Compression Distance (NCD) metric. Results indicate that a near two-fold increase in accuracy can be achieved for NCD by applying simple preprocessing techniques. Nonetheless, the cosine similarity metric still significantly outperforms NCD. Finally, we show that using our Wikipedia-based method to augment the cosine vector space model provides superior results to either in isolation. Combining the two methods leads to an increased correlation of Pearson p = 0:72 over the Lee (2005) document similarity dataset, which matches the reported result for the state-of-the-art Explicit Semantic Analysis (ESA) technique, while requiring less than 10% of the Wikipedia database as required by ESA. As a use case for document similarity techniques, a purely content-based news-article recommender system is designed and implemented for a large online media company. This system is used to gather additional human-generated relevance ratings which we use to evaluate the performance of three state-of-the-art document similarity metrics for providing content-based document recommendations.
AFRIKAANSE OPSOMMING: Die Wêreldwye-Web het ’n vlak van inligting-oorbelading tot gevolg gehad soos nog nooit tevore. Rekenaars is baie effektief met die verwerking en groepering van numeriese en binêre data, maar die konsepsuele groepering van natuurlike-taal data is aansienlik moeiliker om te outomatiseer. Tradisioneel berus sulke algoritmes op eenvoudige sleutelwoordherkenningstegnieke of waarskynlikheidsmetodes om semantiese verwantskappe te bereken, maar hierdie benaderings modelleer nie konsepsuele verwantskappe, soos gemeet deur die mens, baie akkuraat nie. In hierdie tesis stel ons die gebruik van ’n nuwe aktiverings-verspreidingstrategie (AV) voor waarmee inter-konsep verwantskappe bereken kan word, deur die artikel skakelstruktuur van Wikipedia te modelleer as ’n assosiatiewe netwerk. Die AV tegniek word aangepas om te funksioneer oor die Wikipedia skakelstruktuur, en verskeie probleme wat hiermee gepaard gaan word aangespreek. Inter-konsep en inter-dokument verwantskapsmaatstawwe word ontwikkel wat gebruik maak van AV om die konsepsuele verwantskap tussen twee konsepte en twee natuurlike-taal dokumente te bereken. Ons evalueer hierdie benadering oor twee dokument-verwantskap datastelle en die resultate vergelyk goed met die van ander toonaangewende metodes. Verder word teks-voorverwerkingstegnieke ondersoek in terme van die moontlike verbetering wat dit tot gevolg kan hê op die werksverrigting van die bekende kosinus vektorruimtemaatstaf en die genormaliseerde kompressie-afstandmaatstaf (GKA). Resultate dui daarop dat GKA se akkuraatheid byna verdubbel kan word deur gebruik te maak van eenvoudige voorverwerkingstegnieke, maar dat die kosinus vektorruimtemaatstaf steeds aansienlike beter resultate lewer. Laastens wys ons dat die Wikipedia-gebasseerde metode gebruik kan word om die vektorruimtemaatstaf aan te vul tot ’n gekombineerde maatstaf wat beter resultate lewer as enige van die twee metodes afsonderlik. Deur die twee metodes te kombineer lei tot ’n verhoogde korrelasie van Pearson p = 0:72 oor die Lee dokument-verwantskap datastel. Dit is gelyk aan die gerapporteerde resultaat vir Explicit Semantic Analysis (ESA), die huidige beste Wikipedia-gebasseerde tegniek. Ons benadering benodig egter minder as 10% van die Wikipedia databasis wat benodig word vir ESA. As ’n toetstoepassing vir dokument-verwantskaptegnieke ontwerp en implementeer ons ’n stelsel vir ’n aanlyn media-maatskappy wat nuusartikels aanbeveel vir gebruikers, slegs op grond van die artikels se inhoud. Joernaliste wat die stelsel gebruik ken ’n punt toe aan elke aanbeveling en ons gebruik hierdie data om die akkuraatheid van drie toonaangewende maatstawwe vir dokument-verwantskap te evalueer in die konteks van inhoud-gebasseerde nuus-artikel aanbevelings.
21
Laurier, Cyril François. "Automatic Classification of musical mood by content-based analysis." Doctoral thesis, Universitat Pompeu Fabra, 2011. http://hdl.handle.net/10803/51582.
Full textAbstract:
In this work, we focus on automatically classifying music by mood. For this purpose, we propose computational models using information extracted from the audio signal. The foundations of such algorithms are based on techniques from signal processing, machine learning and information retrieval. First, by studying the tagging behavior of a music social network, we find a model to represent mood. Then, we propose a method for automatic music mood classification. We analyze the contributions of audio descriptors and how their values are related to the observed mood. We also propose a multimodal version using lyrics, contributing to the field of text retrieval. Moreover, after showing the relation between mood and genre, we present a new approach using automatic music genre classification. We demonstrate that genre-based mood classifiers give higher accuracies than standard audio models. Finally, we propose a rule extraction technique to explicit our models.En esta tesis, nos centramos en la clasificación automática de música a partir de la detección de la emoción que comunica. Primero, estudiamos cómo los miembros de una red social utilizan etiquetas y palabras clave para describir la música y las emociones que evoca, y encontramos un modelo para representar los estados de ánimo. Luego, proponemos un método de clasificación automática de emociones. Analizamos las contribuciones de descriptores de audio y cómo sus valores están relacionados con los estados de ánimo. Proponemos también una versión multimodal de nuestro algoritmo, usando las letras de canciones. Finalmente, después de estudiar la relación entre el estado de ánimo y el género musical, presentamos un método usando la clasificación automática por género. A modo de recapitulación conceptual y algorítmica, proponemos una técnica de extracción de reglas para entender como los algoritmos de aprendizaje automático predicen la emoción evocada por la música
22
Wilhelm, Thomas. "Entwurf und Implementierung eines Frameworks zur Analyse und Evaluation von Verfahren im Information Retrieval." Master's thesis, [S.l. : s.n.], 2008. https://monarch.qucosa.de/id/qucosa%3A18962.
Full text
23
Jeong, Ki Tai. "A Common Representation Format for Multimedia Documents." Thesis, University of North Texas, 2002. https://digital.library.unt.edu/ark:/67531/metadc3336/.
Full textAbstract:
Multimedia documents are composed of multiple file format combinations, such as image and text, image and sound, or image, text and sound. The type of multimedia document determines the form of analysis for knowledge architecture design and retrieval methods. Over the last few decades, theories of text analysis have been proposed and applied effectively. In recent years, theories of image and sound analysis have been proposed to work with text retrieval systems and progressed quickly due in part to rapid progress in computer processing speed. Retrieval of multimedia documents formerly was divided into the categories of image and text, and image and sound. While standard retrieval process begins from text only, methods are developing that allow the retrieval process to be accomplished simultaneously using text and image. Although image processing for feature extraction and text processing for term extractions are well understood, there are no prior methods that can combine these two features into a single data structure. This dissertation will introduce a common representation format for multimedia documents (CRFMD) composed of both images and text. For image and text analysis, two techniques are used: the Lorenz Information Measurement and the Word Code. A new process named Jeong's Transform is demonstrated for extraction of text and image features, combining the two previous measurements to form a single data structure. Finally, this single data measurements to form a single data structure. Finally, this single data structure is analyzed by using multi-dimensional scaling. This allows multimedia objects to be represented on a two-dimensional graph as vectors. The distance between vectors represents the magnitude of the difference between multimedia documents. This study shows that image classification on a given test set is dramatically improved when text features are encoded together with image features. This effect appears to hold true even when the available text is diffused and is not uniform with the image features. This retrieval system works by representing a multimedia document as a single data structure. CRFMD is applicable to other areas of multimedia document retrieval and processing, such as medical image retrieval, World Wide Web searching, and museum collection retrieval.
24
Xu, Xiaoqian. "Shape Matching, Relevance Feedback, and Indexing with Application to Spine X-Ray Image Retrieval." Diss., CLICK HERE for online access, 2006. http://contentdm.lib.byu.edu/ETD/image/etd1648.pdf.
Full text
25
Lunze, Torsten. "Recommendation in Enterprise 2.0 Social Media Streams." Doctoral thesis, Saechsische Landesbibliothek- Staats- und Universitaetsbibliothek Dresden, 2014. http://nbn-resolving.de/urn:nbn:de:bsz:14-qucosa-153603.
Full textAbstract:
A social media stream allows users to share user-generated content as well as aggregate different external sources into one single stream. In Enterprise 2.0 such social media streams empower co-workers to share their information and to work efficiently and effectively together while replacing email communication. As more users share information it becomes impossible to read the complete stream leading to an information overload. Therefore, it is crucial to provide the users a personalized stream that suggests important and unread messages. The main characteristic of an Enterprise 2.0 social media stream is that co-workers work together on projects represented by topics: the stream is topic-centered and not user-centered as in public streams such as Facebook or Twitter.
A lot of work has been done dealing with recommendation in a stream or for news recommendation. However, none of the current research approaches deal with the characteristics of an Enterprise 2.0 social media stream to recommend messages. The existing systems described in the research mainly deal with news recommendation for public streams and lack the applicability for Enterprise 2.0 social media streams.
In this thesis a recommender concept is developed that allows the recommendation of messages in an Enterprise 2.0 social media stream. The basic idea is to extract features from a new message and use those features to compute a relevance score for a user. Additionally, those features are used to learn a user model and then use the user model for scoring new messages. This idea works without using explicit user feedback and assures a high user acceptance because no intense rating of messages is necessary. With this idea a content-based and collaborative-based approach is developed. To reflect the topic-centered streams a topic-specific user model is introduced which learns a user model independently for each topic.
There are constantly new terms that occur in the stream of messages. For improving the quality of the recommendation (by finding more relevant messages) the recommender should be able to handle the new terms. Therefore, an approach is developed which adapts a user model if unknown terms occur by using terms of similar users or topics. Also, a short- and long-term approach is developed which tries to detect short-term interests of users. Only if the interest of a user occurs repeatedly over a certain time span are terms transferred to the long-term user model.
The approaches are evaluated against a dataset obtained through an Enterprise 2.0 social media stream application. The evaluation shows the overall applicability of the concept. Specifically the evaluation shows that a topic-specific user model outperforms a global user model and also that adapting the user model according to similar users leads to an increase in the quality of the recommendation. Interestingly, the collaborative-based approach cannot reach the quality of the content-based approach.
26
Tarczyńska, Anna. "Methods of Text Information Extraction in Digital Videos." Thesis, Blekinge Tekniska Högskola, Sektionen för datavetenskap och kommunikation, 2012. http://urn.kb.se/resolve?urn=urn:nbn:se:bth-2656.
Full textAbstract:
Context The huge amount of existing digital video files needs to provide indexing to make it available for customers (easier searching). The indexing can be provided by text information extraction. In this thesis we have analysed and compared methods of text information extraction in digital videos. Furthermore, we have evaluated them in the new context proposed by us, namely usefulness in sports news indexing and information retrieval. Objectives The objectives of this thesis are as follows: providing a better understanding of the nature of text extraction; performing a systematic literature review on various methods of text information extraction in digital videos of TV sports news; designing and executing an experiment in the testing environment; evaluating available and promising methods of text information extraction from digital video files in the proposed context associated with video sports news indexing and retrieval; providing an adequate solution in the proposed context described above. Methods This thesis consists of three research methods: Systematic Literature Review, Video Content Analysis with the checklist, and Experiment. The Systematic Literature Review has been used to study the nature of text information extraction, to establish the methods and challenges, and to specify the effective way of conducting the experiment. The video content analysis has been used to establish the context for the experiment. Finally, the experiment has been conducted to answer the main research question: How useful are the methods of text information extraction for indexation of video sports news and information retrieval? Results Through the Systematic Literature Review we identified 29 challenges of the text information extraction methods, and 10 chains between them. We extracted 21 tools and 105 different methods, and analyzed the relations between them. Through Video Content Analysis we specified three groups of probability of text extraction from video, and 14 categories for providing video sports news indexation with the taxonomy hierarchy. We have conducted the Experiment on three videos files, with 127 frames, 8970 characters, and 1814 words, using the only available MoCA tool. As a result, we reported 10 errors and proposed recommendations for each of them. We evaluated the tool according to the categories mentioned above and offered four advantages, and nine disadvantages of the Tool mentioned above. Conclusions It is hard to compare the methods described in the literature, because the tools are not available for testing, and they are not compared with each other. Furthermore, the values of recall and precision measures highly depend on the quality of the text contained in the video. Therefore, performing the experiments on the same indexed database is necessary. However, the text information extraction is time consuming (because of huge amount of frames in video), and even high character recognition rate gives low word recognition rate. Therefore, the usefulness of text information extraction for video indexation is still low. Because most of the text information contained in the videos news is inserted in post-processing, the text extraction could be provided in the root: during the processing of the original video, by the broadcasting company (e.g. by automatically saving inserted text in separate file). Then the text information extraction will not be necessary for managing the new video filesThe huge amount of existing digital video files needs to provide indexing to make it available for customers (easier searching). The indexing can be provided by text information extraction. In this thesis we have analysed and compared methods of text information extraction in digital videos. Furthermore, we have evaluated them in the new context proposed by us, namely usefulness in sports news indexing and information retrieval.
27
Le, Huu Ton. "Improving image representation using image saliency and information gain." Thesis, Poitiers, 2015. http://www.theses.fr/2015POIT2287/document.
Full textAbstract:
De nos jours, avec le développement des nouvelles technologies multimédia, la recherche d’images basée sur le contenu visuel est un sujet de recherche en plein essor avec de nombreux domaines d'application: indexation et recherche d’images, la graphologie, la détection et le suivi d’objets... Un des modèles les plus utilisés dans ce domaine est le sac de mots visuels qui tire son inspiration de la recherche d’information dans des documents textuels. Dans ce modèle, les images sont représentées par des histogrammes de mots visuels à partir d'un dictionnaire visuel de référence. La signature d’une image joue un rôle important car elle détermine la précision des résultats retournés par le système de recherche.Dans cette thèse, nous étudions les différentes approches concernant la représentation des images. Notre première contribution est de proposer une nouvelle méthodologie pour la construction du vocabulaire visuel en utilisant le gain d'information extrait des mots visuels. Ce gain d’information est la combinaison d’un modèle de recherche d’information avec un modèle d'attention visuelle.Ensuite, nous utilisons un modèle d'attention visuelle pour améliorer la performance de notre modèle de sacs de mots visuels. Cette étude de la saillance des descripteurs locaux souligne l’importance d’utiliser un modèle d’attention visuelle pour la description d’une image.La dernière contribution de cette thèse au domaine de la recherche d’information multimédia démontre comment notre méthodologie améliore le modèle des sacs de phrases visuelles. Finalement, une technique d’expansion de requêtes est utilisée pour augmenter la performance de la recherche par les deux modèles étudiésNowadays, along with the development of multimedia technology, content based image retrieval (CBIR) has become an interesting and active research topic with an increasing number of application domains: image indexing and retrieval, face recognition, event detection, hand writing scanning, objects detection and tracking, image classification, landmark detection... One of the most popular models in CBIR is Bag of Visual Words (BoVW) which is inspired by Bag of Words model from Information Retrieval field. In BoVW model, images are represented by histograms of visual words from a visual vocabulary. By comparing the images signatures, we can tell the difference between images. Image representation plays an important role in a CBIR system as it determines the precision of the retrieval results.In this thesis, image representation problem is addressed. Our first contribution is to propose a new framework for visual vocabulary construction using information gain (IG) values. The IG values are computed by a weighting scheme combined with a visual attention model. Secondly, we propose to use visual attention model to improve the performance of the proposed BoVW model. This contribution addresses the importance of saliency key-points in the images by a study on the saliency of local feature detectors. Inspired from the results from this study, we use saliency as a weighting or an additional histogram for image representation.The last contribution of this thesis to CBIR shows how our framework enhances the BoVP model. Finally, a query expansion technique is employed to increase the retrieval scores on both BoVW and BoVP models
28
Bergqvist, Martin, and Jim Glansk. "Fördelar med att applicera Collaborative Filtering på Steam : En utforskande studie." Thesis, Högskolan i Borås, Akademin för bibliotek, information, pedagogik och IT, 2018. http://urn.kb.se/resolve?urn=urn:nbn:se:hb:diva-14129.
Full textAbstract:
Rekommendationssystem används överallt. På populära plattformar såsom Netflix och Amazon får du alltid rekommendationer på vad som är nästa lämpliga film eller inköp, baserat på din personliga profil. Detta sker genom korsreferering mellan användare och produkter för att finna sannolika mönster. Syftet med studien har varit att jämföra de två prevalenta tillvägagångssätten att skapa rekommendationer, på en annorlunda datamängd, där ”best practice” inte nödvändigtvis är tillämpbart. Som följd därav, har jämförelse gjorts på effektiviteten av Content-based Filtering kontra Collaborative Filtering, på Steams spelplattform, i syfte att etablera potential för en bättre lösning. Detta angreps genom att samla in data från Steam; Bygga en Content-based Filtering motor som baslinje för att representera Steams nuvarande rekommendationssystem, samt en motsvarande Collaborative Filtering motor, baserad på en standard-implementation, att jämföra mot. Under studiens gång visade det sig att Content-based Filtering prestanda initiellt växte linjärt medan spelarbasen på ett givet spel ökade. Collaborative Filtering däremot hade en exponentiell prestationskurva för spel med få spelare, för att sedan plana ut på en nivå som prestationsmässigt överträffade jämförelsmetoden. Den praktiska signifikansen av dessa resultat torde rättfärdiga en mer utbredd implementering av Collaborative Filtering även där man normalt avstår till förmån för Content-based Filtering då det är enklare att implementera och ger acceptabla resultat. Då våra resultat visar på såpass stor avvikelse redan vid basmodeller, är det här en attityd som mycket väl kan förändras. Collaborative Filtering har varit sparsamt använt på mer mångfacetterade datamängder, men våra resultat visar på potential att överträffa Content-based Filtering med relativt liten insats även på sådana datamängder. Detta kan gynna alla inköps- och community-kombinerade plattformar, då det finns möjlighet att övervaka användandet av inköpen i realtid, vilket möjliggör för justeringar av de faktorer som kan visa sig resultera i felrepresentation.The use of recommender systems is everywhere. On popular platforms such as Netflix and Amazon, you are always given new recommendations on what to consume next, based on your specific profiling. This is done by cross-referencing users and products to find probable patterns. The aims of this study were to compare the two main ways of generating recommendations, in an unorthodox dataset where “best practice” might not apply. Subsequently, recommendation efficiency was compared between Content Based Filtering and Collaborative Filtering, on the gaming-platform of Steam, in order to establish if there was potential for a better solution. We approached this by gathering data from Steam, building a representational baseline Content-based Filtering recommendation-engine based on what is currently used by Steam, and a competing Collaborative Filtering engine based on a standard implementation. In the course of this study, we found that while Content-based Filtering performance initially grew linearly as the player base of a game increased, Collaborative Filtering’s performance grew exponentially from a small player base, to plateau at a performance-level exceeding the comparison. The practical consequence of these findings would be the justification to apply Collaborative Filtering even on smaller, more complex sets of data than is normally done; The justification being that Content-based Filtering is easier to implement and yields decent results. With our findings showing such a big discrepancy even at basic models, this attitude might well change. The usage of Collaborative Filtering has been used scarcely on the more multifaceted datasets, but our results show that the potential to exceed Content-based Filtering is rather easily obtainable on such sets as well. This potentially benefits all purchase/community-combined platforms, as the usage of the purchase is monitorable on-line, and allows for the adjustments of misrepresentational factors as they appear.
29
Celma, Herrada Òscar. "Music recommendation and discovery in the long tail." Doctoral thesis, Universitat Pompeu Fabra, 2009. http://hdl.handle.net/10803/7557.
Full textAbstract:
Avui en dia, la música està esbiaixada cap al consum d'alguns artistes molt populars. Per exemple, el 2007 només l'1% de totes les cançons en format digital va representar el 80% de les vendes. De la mateixa manera, només 1.000 àlbums varen representar el 50% de totes les vendes, i el 80% de tots els àlbums venuts es varen comprar menys de 100 vegades. Es clar que hi ha una necessitat per tal d'ajudar a les persones a filtrar, descobrir, personalitzar i recomanar música, a partir de l'enorme quantitat de contingut musical disponible. Els algorismes de recomanació de música actuals intenten predir amb precisió el que els usuaris demanen escoltar. Tanmateix, molt sovint aquests algoritmes tendeixen a recomanar artistes famosos, o coneguts d'avantmà per l'usuari. Això fa que disminueixi l'eficàcia i utilitat de les recomanacions, ja que aquests algorismes es centren bàsicament en millorar la precisió de les recomanacions. És a dir, tracten de fer prediccions exactes sobre el que un usuari pugui escoltar o comprar, independentment de quant útils siguin les recomanacions generades. En aquesta tesi destaquem la importància que l'usuari valori les recomanacions rebudes. Per aquesta raó modelem la corba de popularitat dels artistes, per tal de poder recomanar música interessant i desconeguda per l'usuari.
Les principals contribucions d'aquesta tesi són: (i) un nou enfocament basat en l'anàlisi de xarxes complexes i la popularitat dels productes, aplicada als sistemes de recomanació, (ii) una avaluació centrada en l'usuari, que mesura la importància i la desconeixença de les recomanacions, i (iii) dos prototips que implementen la idees derivades de la tasca teòrica. Els resultats obtinguts tenen una clara implicació per aquells sistemes de recomanació que ajuden a l'usuari a explorar i descobrir continguts que els pugui agradar.
Actualmente, el consumo de música está sesgada hacia algunos artistas muy populares. Por ejemplo, en el año 2007 sólo el 1% de todas las canciones en formato digital representaron el 80% de las ventas. De igual modo, únicamente 1.000 álbumes representaron el 50% de todas las ventas, y el 80% de todos los álbumes vendidos se compraron menos de 100 veces. Existe, pues, una necesidad de ayudar a los usuarios a filtrar, descubrir, personalizar y recomendar música a partir de la enorme cantidad de contenido musical existente. Los algoritmos de recomendación musical existentes intentan predecir con precisión lo que la gente quiere escuchar. Sin embargo, muy a menudo estos algoritmos tienden a recomendar o bien artistas famosos, o bien artistas ya conocidos de antemano por el usuario.Esto disminuye la eficacia y la utilidad de las recomendaciones, ya que estos algoritmos se centran en mejorar la precisión de las recomendaciones. Con lo cuál, tratan de predecir lo que un usuario pudiera escuchar o comprar, independientemente de lo útiles que sean las recomendaciones generadas.
En este sentido, la tesis destaca la importancia de que el usuario valore las recomendaciones propuestas. Para ello, modelamos la curva de popularidad de los artistas con el fin de recomendar música interesante y, a la vez, desconocida para el usuario.Las principales contribuciones de esta tesis son: (i) un nuevo enfoque basado en el análisis de redes complejas y la popularidad de los productos, aplicada a los sistemas de recomendación,(ii) una evaluación centrada en el usuario que mide la calidad y la novedad de las recomendaciones, y (iii) dos prototipos que implementan las ideas derivadas de la labor teórica. Los resultados obtenidos tienen importantes implicaciones para los sistemas de recomendación que ayudan al usuario a explorar y descubrir contenidos que le puedan gustar.
Music consumption is biased towards a few popular artists. For instance, in 2007 only 1% of all digital tracks accounted for 80% of all sales. Similarly, 1,000 albums accounted for 50% of all album sales, and 80% of all albums sold were purchased less than 100 times. There is a need to assist people to filter, discover, personalise and recommend from the huge amount of music content available along the Long Tail.Current music recommendation algorithms try to accurately predict what people demand to listen to. However, quite often these algorithms tend to recommend popular -or well-known to the user- music, decreasing the effectiveness of the recommendations. These approaches focus on improving the accuracy of the recommendations. That is, try to make accurate predictions about what a user could listen to, or buy next, independently of how useful to the user could be the provided recommendations.
In this Thesis we stress the importance of the user's perceived quality of the recommendations. We model the Long Tail curve of artist popularity to predict -potentially- interesting and unknown music, hidden in the tail of the popularity curve. Effective recommendation systems should promote novel and relevant material (non-obvious recommendations), taken primarily from the tail of a popularity distribution.
The main contributions of this Thesis are: (i) a novel network-based approach for recommender systems, based on the analysis of the item (or user) similarity graph, and the popularity of the items, (ii) a user-centric evaluation that measures the user's relevance and novelty of the recommendations, and (iii) two prototype systems that implement the ideas derived from the theoretical work. Our findings have significant implications for recommender systems that assist users to explore the Long Tail, digging for content they might like.
30
Beecks, Christian [Verfasser]. "Distance-based similarity models for content-based multimedia retrieval / Christian Beecks." Aachen : Hochschulbibliothek der Rheinisch-Westfälischen Technischen Hochschule Aachen, 2013. http://d-nb.info/1046647245/34.
Full text
31
Haro, Berois Martín. "Statistical distribution of common audio features : encounters in a heavy-tailed universe." Doctoral thesis, Universitat Pompeu Fabra, 2013. http://hdl.handle.net/10803/128623.
Full textAbstract:
In the last few years some Music Information Retrieval (MIR) researchers have spotted important drawbacks in applying standard successful-in-monophonic algorithms to polyphonic music classification and similarity assessment. Noticeably, these so called “Bag-of-Frames” (BoF) algorithms share a common set of assumptions. These assumptions are substantiated in the belief that the numerical descriptions extracted from short-time audio excerpts (or frames) are enough to capture relevant information for the task at hand, that these frame-based audio descriptors are time independent, and that descriptor frames are well described by Gaussian statistics. Thus, if we want to improve current BoF algorithms we could: i) improve current audio descriptors, ii) include temporal information within algorithms working with polyphonic music, and iii) study and characterize the real statistical properties of these frame-based audio descriptors. From a literature review, we have detected that many works focus on the first two improvements, but surprisingly, there is a lack of research in the third one. Therefore, in this thesis we analyze and characterize the statistical distribution of common audio descriptors of timbre, tonal and loudness information. Contrary to what is usually assumed, our work shows that the studied descriptors are heavy-tailed distributed and thus, they do not belong to a Gaussian universe. This new knowledge led us to propose new algorithms that show improvements over the BoF approach in current MIR tasks such as genre classification, instrument detection, and automatic tagging of music. Furthermore, we also address new MIR tasks such as measuring the temporal evolution of Western popular music. Finally, we highlight some promising paths for future audio-content MIR research that will inhabit a heavy-tailed universe.En el campo de la extracción de información musical o Music Information Retrieval (MIR), los algoritmos llamados Bag-of-Frames (BoF) han sido aplicados con éxito en la clasificación y evaluación de similitud de señales de audio monofónicas. Por otra parte, investigaciones recientes han señalado problemas importantes a la hora de aplicar dichos algoritmos a señales de música polifónica. Estos algoritmos suponen que las descripciones numéricas extraídas de los fragmentos de audio de corta duración (o frames ) son capaces de capturar la información necesaria para la realización de las tareas planteadas, que el orden temporal de estos fragmentos de audio es irrelevante y que las descripciones extraídas de los segmentos de audio pueden ser correctamente descritas usando estadísticas Gaussianas. Por lo tanto, si se pretende mejorar los algoritmos BoF actuales se podría intentar: i) mejorar los descriptores de audio, ii) incluir información temporal en los algoritmos que trabajan con música polifónica y iii) estudiar y caracterizar las propiedades estadísticas reales de los descriptores de audio. La bibliografía actual sobre el tema refleja la existencia de un número considerable de trabajos centrados en las dos primeras opciones de mejora, pero sorprendentemente, hay una carencia de trabajos de investigación focalizados en la tercera opción. Por lo tanto, esta tesis se centra en el análisis y caracterización de la distribución estadística de descriptores de audio comúnmente utilizados para representar información tímbrica, tonal y de volumen. Al contrario de lo que se asume habitualmente, nuestro trabajo muestra que los descriptores de audio estudiados se distribuyen de acuerdo a una distribución de “cola pesada” y por lo tanto no pertenecen a un universo Gaussiano. Este descubrimiento nos permite proponer nuevos algoritmos que evidencian mejoras importantes sobre los algoritmos BoF actualmente utilizados en diversas tareas de MIR tales como clasificación de género, detección de instrumentos musicales y etiquetado automático de música. También nos permite proponer nuevas tareas tales como la medición de la evolución temporal de la música popular occidental. Finalmente, presentamos algunas prometedoras líneas de investigación para tareas de MIR ubicadas, a partir de ahora, en un universo de “cola pesada”.
En l’àmbit de la extracció de la informació musical o Music Information Retrieval (MIR), els algorismes anomenats Bag-of-Frames (BoF) han estat aplicats amb èxit en la classificació i avaluació de similitud entre senyals monofòniques. D’altra banda, investigacions recents han assenyalat importants inconvenients a l’hora d’aplicar aquests mateixos algorismes en senyals de música polifònica. Aquests algorismes BoF suposen que les descripcions numèriques extretes dels fragments d’àudio de curta durada (frames) son suficients per capturar la informació rellevant per als algorismes, que els descriptors basats en els fragments son independents del temps i que l’estadística Gaussiana descriu correctament aquests descriptors. Per a millorar els algorismes BoF actuals doncs, es poden i) millorar els descriptors, ii) incorporar informació temporal dins els algorismes que treballen amb música polifònica i iii) estudiar i caracteritzar les propietats estadístiques reals d’aquests descriptors basats en fragments d’àudio. Sorprenentment, de la revisió bibliogràfica es desprèn que la majoria d’investigacions s’han centrat en els dos primers punts de millora mentre que hi ha una mancança quant a la recerca en l’àmbit del tercer punt. És per això que en aquesta tesi, s’analitza i caracteritza la distribució estadística dels descriptors més comuns de timbre, to i volum. El nostre treball mostra que contràriament al què s’assumeix, els descriptors no pertanyen a l’univers Gaussià sinó que es distribueixen segons una distribució de “cua pesada”. Aquest descobriment ens permet proposar nous algorismes que evidencien millores importants sobre els algorismes BoF utilitzats actualment en diferents tasques com la classificació del gènere, la detecció d’instruments musicals i l’etiquetatge automàtic de música. Ens permet també proposar noves tasques com la mesura de l’evolució temporal de la música popular occidental. Finalment, presentem algunes prometedores línies d’investigació per a tasques de MIR ubicades a partir d’ara en un univers de “cua pesada”.
32
Hernández, Mesa Pilar [Verfasser]. "Design and analysis of a content-based image retrieval system / Pilar Hernández Mesa." Karlsruhe : KIT Scientific Publishing, 2017. http://www.ksp.kit.edu.
Full text
33
Désoyer, Adèle. "Appariement de contenus textuels dans le domaine de la presse en ligne : développement et adaptation d'un système de recherche d'information." Thesis, Paris 10, 2017. http://www.theses.fr/2017PA100119/document.
Full textAbstract:
L'objectif de cette thèse, menée dans un cadre industriel, est d'apparier des contenus textuels médiatiques. Plus précisément, il s'agit d'apparier à des articles de presse en ligne des vidéos pertinentes, pour lesquelles nous disposons d'une description textuelle. Notre problématique relève donc exclusivement de l'analyse de matériaux textuels, et ne fait intervenir aucune analyse d'image ni de langue orale. Surviennent alors des questions relatives à la façon de comparer des objets textuels, ainsi qu'aux critères mobilisés pour estimer leur degré de similarité. L'un de ces éléments est selon nous la similarité thématique de leurs contenus, autrement dit le fait que deux documents doivent relater le même sujet pour former une paire pertinente. Ces problématiques relèvent du domaine de la recherche d'information (ri), dans lequel nous nous ancrons principalement. Par ailleurs, lorsque l'on traite des contenus d'actualité, la dimension temporelle est aussi primordiale et les problématiques qui l'entourent relèvent de travaux ayant trait au domaine du topic detection and tracking (tdt) dans lequel nous nous inscrivons également.Le système d'appariement développé dans cette thèse distingue donc différentes étapes qui se complètent. Dans un premier temps, l'indexation des contenus fait appel à des méthodes de traitement automatique des langues (tal) pour dépasser la représentation classique des textes en sac de mots. Ensuite, deux scores sont calculés pour rendre compte du degré de similarité entre deux contenus : l'un relatif à leur similarité thématique, basé sur un modèle vectoriel de ri; l'autre à leur proximité temporelle, basé sur une fonction empirique. Finalement, un modèle de classification appris à partir de paires de documents, décrites par ces deux scores et annotées manuellement, permet d'ordonnancer les résultats.L'évaluation des performances du système a elle aussi fait l'objet de questionnements dans ces travaux de thèse. Les contraintes imposées par les données traitées et le besoin particulier de l'entreprise partenaire nous ont en effet contraints à adopter une alternative au protocole classique d'évaluation en ri, le paradigme de CranfieldThe goal of this thesis, conducted within an industrial framework, is to pair textual media content. Specifically, the aim is to pair on-line news articles to relevant videos for which we have a textual description. The main issue is then a matter of textual analysis, no image or spoken language analysis was undertaken in the present study. The question that arises is how to compare these particular objects, the texts, and also what criteria to use in order to estimate their degree of similarity. We consider that one of these criteria is the topic similarity of their content, in other words, the fact that two documents have to deal with the same topic to form a relevant pair. This problem fall within the field of information retrieval (ir) which is the main strategy called upon in this research. Furthermore, when dealing with news content, the time dimension is of prime importance. To address this aspect, the field of topic detection and tracking (tdt) will also be explored.The pairing system developed in this thesis distinguishes different steps which complement one another. In the first step, the system uses natural language processing (nlp) methods to index both articles and videos, in order to overcome the traditionnal bag-of-words representation of texts. In the second step, two scores are calculated for an article-video pair: the first one reflects their topical similarity and is based on a vector space model; the second one expresses their proximity in time, based on an empirical function. At the end of the algorithm, a classification model learned from manually annotated document pairs is used to rank the results.Evaluation of the system's performances raised some further questions in this doctoral research. The constraints imposed both by the data and the specific need of the partner company led us to adapt the evaluation protocol traditionnal used in ir, namely the cranfield paradigm. We therefore propose an alternative solution for evaluating the system that takes all our constraints into account
34
Kaufman, Jaime C. "A Hybrid Approach to Music Recommendation: Exploiting Collaborative Music Tags and Acoustic Features." UNF Digital Commons, 2014. http://digitalcommons.unf.edu/etd/540.
Full textAbstract:
Recommendation systems make it easier for an individual to navigate through large datasets by recommending information relevant to the user. Companies such as Facebook, LinkedIn, Twitter, Netflix, Amazon, Pandora, and others utilize these types of systems in order to increase revenue by providing personalized recommendations. Recommendation systems generally use one of the two techniques: collaborative filtering (i.e., collective intelligence) and content-based filtering.
Systems using collaborative filtering recommend items based on a community of users, their preferences, and their browsing or shopping behavior. Examples include Netflix, Amazon shopping, and Last.fm. This approach has been proven effective due to increased popularity, and its accuracy improves as its pool of users expands. However, the weakness with this approach is the Cold Start problem. It is difficult to recommend items that are either brand new or have no user activity.
Systems that use content-based filtering recommend items based on extracted information from the actual content. A popular example of this approach is Pandora Internet Radio. This approach overcomes the Cold Start problem. However, the main issue with this approach is its heavy demand on computational power. Also, the semantic meaning of an item may not be taken into account when producing recommendations.
In this thesis, a hybrid approach is proposed by utilizing the strengths of both collaborative and content-based filtering techniques. As proof-of-concept, a hybrid music recommendation system was developed and evaluated by users. The results show that this system effectively tackles the Cold Start problem and provides more variation on what is recommended.
35
Souza, Juliana Pereira de. "Modelo de qualidade para o desenvolvimento e avaliação da viabilidade clínica de sistemas de recuperação de imagens médicas baseadas em conteúdo." Universidade de São Paulo, 2012. http://www.teses.usp.br/teses/disponiveis/82/82131/tde-15022013-142541/.
Full textAbstract:
Com a crescente utilização de imagens médicas na prática clínica, torna-se necessária a introdução de tecnologias que garantam o armazenamento, indexação e recuperação eficaz dessas imagens. O sistema de recuperação de imagens médicas baseada em conteúdo (S-CBIR) compõe a base de tecnologias computacionais que oferecem aos usuários médicos aplicativos para apoio ao diagnóstico, sendo capaz de responder a consultas por similaridade por meio de características pictóricas extraídas das imagens médicas. Embora as pesquisas em S-CBIR tenham iniciado há quase duas décadas, atualmente existe uma discrepância em relação à quantidade de trabalhos publicados na literatura e os sistemas que, de fato, foram implementados e avaliados. Além disso, muitos protótipos vêm sendo discutidos, mas até o final da escrita desta tese, não foram encontradas evidências de que algum deles esteja disponível comercialmente. Essa limitação é conhecida pela comunidade científica da área por gap de aplicação. Em geral, isso ocorre devido à dificuldade dessas aplicações em superar alguns desafios, como a divergência entre os resultados obtidos automaticamente pelo sistema e aqueles esperados pelos médicos (gap semântico), entre outros gap. Outros fatores também podem ser relatados, como a tendência da não utilização de modelos de qualidade sistematizados para o desenvolvimento dos sistemas, e a carência de modelos que sejam específicos no domínio de aplicação. Com base nesses desafios e em boas práticas de métodos, técnicas e ferramentas da Engenharia de Software, esta tese apresenta um Modelo de Qualidade para melhorias de S-CBIR (MQ-SCBIR), que tem por objetivo apoiar o desenvolvimento e avaliação de S-CBIR, a partir de diretrizes para aumentar o nível de qualidade, buscando a superação do gap de aplicação. O MQ-SCBIR foi construído com base em: evidências adquiridas por meio de uma revisão sistemática e pesquisa empírica sobre como esses sistemas vêm sendo desenvolvidos e avaliados na literatura e na prática; resultados da avaliação de um S-CBIR baseados em testes heurísticos em um ambiente real; modelos bem estabelecidos, como o Capability Maturity Model Integration e Melhoria de Processo do Software Brasileiro; e em experiências pessoais. O uso do MQ-SCBIR pode trazer benefícios para as organizações desenvolvedoras, como a redução da complexidade no desenvolvimento, incluindo a garantia de implementação de boas práticas de qualidade de software e práticas específicas para a superação das limitações de S-CBIR durante o processo de desenvolvimento.The development of technologies for storing, indexing and recovering clinical images is paramount to support the increasing use of these images in clinical diagnostic evaluation. Content-based image retrieval systems (CBIR-S) are some of the main computational technologies which offer physicians different applications to aid diagnostic processes. They allow similarity queries by extracting pictorial features from medical images. Even though research on S-CBIR started almost two decades ago, there are discrepancies regarding the amount of studies available in the literature and the number of systems which have actually been implemented and evaluated. Many prototypes have been discussed, but up to the moment this study was completed we found no evidence that any of those systems are either commercially available or being currently used in clinical practice. This limitation is known as application gap. In general, this happens due to the difficulty to overcome some obstacles, such as the differences between the results retrieved automatically by the system and those expected by the physicians (semantic gap). Other factors can also be described, such as the tendency towards not using systematic quality models to develop these systems and the lack of specific models for this domain of application. Based on these challenges and also on best practice methods, techniques and tools from software engineering, this work presents a quality model to improve S-CBIR systems (QM-CBIRS). It strives to tackle limitations during the development process by overcoming the semantic gap. The QM-CBIRS was built upon evidence gathered by means of a systematic review on the state-of-the-art and empiric research on the development and evaluation of these systems. Apart from that, results from the assessment of a CBIR-S based on empiric tests and on diagnostic tasks in radiology and well-established software quality models, such as CMMI and the Brazilian Software Improvement Process are presented. Apart from that, results from the assessment of a CBIR-S based on empiric tests and on diagnostic tasks in radiology and well-established software quality models, such as CMMI and the Brazilian Software Improvement Process are presented. The use of QM-CBIRS might be beneficial to development teams in many ways, for example, by increasing the quality of CBIR systems and reducing complexity, thus surpassing limitations from CBIR systems during the development process.
36
Hsu, Jia-Lien, and 徐嘉連. "Content-based Music Information Retrieval and Analysis." Thesis, 2001. http://ndltd.ncl.edu.tw/handle/33474421178600440765.
Full textAbstract:
博士國立清華大學
資訊工程學系
90
In this thesis, we first discuss the techniques used in content-based music information retrieval. The techniques include the methods to represent music objects, the similarity measures of music objects, and indexing and query processing for music object retrieval. To represent music objects, we introduce three coding schemes, i.e., chord, mubol, and music segment. Various similarity measures are then presented, followed by various index structures and the associated query processing algorithms. The index structures include suffix tree, n-gram, and augmented suffix tree. A qualitative comparison of these techniques is finally performed to show the intrinsic difficulty of the problem of content-based music information retrieval. We also initiate the Ultima project which aims to construct a platform for evaluating various approaches of music information retrieval. Three approaches with the corresponding tree-based, list-based, and (n-gram+tree)-based index structures are implemented. A series of experiments has been carried out. With the support of the experiment results, we compare the performance of index construction and query processing of the three approaches and give a summary for efficient content-based music information retrieval. The feature extraction problem for music objects is also studied to support content-based music information retrieval in searching, classification, recommendation, and so forth. A repeating pattern in music data is defined as a sequence of notes which appears more than once in a music object. The themes are a typical kind of repeating patterns. The themes and other non-trivial repeating patterns are important music features which can be used for both content-based retrieval of music data and music data analysis. We propose two approaches for fast discovering non-trivial repeating patterns in music objects. In the first approach, we develop a data structure called correlative matrix and its associated algorithms for extracting the repeating patterns. In the second approach, we introduce a string-join operation and a data structure called RP-tree for the same purpose. Experiments are performed to compare these two approaches with others. The results are also analyzed to show the efficiency and the effectiveness of our approaches. Further, we extend the problem of finding exact repeating patterns to the one of finding approximate repeating patterns. First, two applications are introduced to motivate our research of finding approximate repeating patterns from sequence data. An approximate repeating pattern is defined as a sequence of symbols which appears more than once under certain approximation types in a data sequence. We define three approximation types, i.e., longer_length, shorter_length, and equal_length. The problems of finding approximate repeating patterns with respect to the three types are specified. By applying the concept of ‘cut’ and ‘pattern_join’ operator, we develop a level-wise approach to solve the problem of finding approximate repeating patterns with respect to the type of longer_length approximation. In addition, we extend the pattern_join operator to the generalized_pattern_join operator for efficiently finding long patterns. The performance study shows that our approach is efficient and also scales well. We also refine our approach to extract repeating patterns from polyphonic music data.
37
Varanguien, de Villepin Audrey. "Content-based color image retrieval." Thesis, 1999. http://hdl.handle.net/1957/33162.
Full textAbstract:
A fully automated method for content-based color image retrieval is developed to extract color and shape content of an image. A color segmentation algorithm based on the k-mean clustering algorithm is used and a saturated distance is proposed to discriminate between two color points in the HSV color space. The feature set describing an image includes main object shape, which is extracted using the morphological operations. The computed image features are tagged within the image and a graphical user interface is presented for retrieving images based on the color and shape of the objects. The experimental results using natural color images demonstrate effectiveness of the proposed method.Graduation date: 2000
38
Huang, Chun-Hong, and 黃俊宏. "Content-Based Information Retrieval on 3D VRML Objects." Thesis, 2003. http://ndltd.ncl.edu.tw/handle/26036273050941537527.
Full textAbstract:
碩士淡江大學
資訊工程學系
91
Because of the interest of 3D images and the popularity of World Wide Web, the number of 3D scene/object and model database throughout the world is growing both in number and in size. VRML (Virtual Reality Modeling Language) is used to model the 3D object of the web page, and become as a standard for 3D information. Most content-based retrieval (CBR) techniques such as shape and color comparison among objects are for image and video. The mechanisms are designed based on 2-D information. In this paper, we propose a new similarity method, based on 3D information extracted from a VRML object database. The system includes multiple subsystems, which will be conducted in two main parts. First part includes: the VRML objects parser system and the normalization system. The second part: includes the comparison system, database management system and a friendly graphical user interface. The method is the first step of our project, which aims to incorporate feature extraction of virtual reality objects, such as chairs, car, and others in 3D space House interior designers can use the proposed system. The user can select proper scenes and furniture in order to meet the requirement of potential customers.
39
Liu, Chueh-Chih, and 劉爵至. "Query By Humming ─ Content-Based MP3 Information Retrieval." Thesis, 2003. http://ndltd.ncl.edu.tw/handle/30994388955937211127.
Full textAbstract:
碩士大同大學
資訊經營研究所
91
In this thesis we investigate the approach of the MP3 content-based retrieval, allowing users to query by humming to take retrieval action. In recent years, Query By Humming (QBH) has become a popular study. Scholars are engrossed in the MIDI music format. Since the prevalence of the Internet and MP3 format appeared, which has the characteristics of little files and the MP3 music quality can complete with that of CD music, raising the revolution of digital music. This thesis focuses on the MP3 music format. The key point is comparing the musical data object by the N-gram approach. The musical data objects are encoded based on a mapping function preprocessed, which integrates Bi-gram, Tri-gram and Four-gram Markov models to look for the best result of proper proportion. Further, the differences of near musical notes proceed to compare the points of similarity. Experiments include initial MP3 music object and MIDI music object test. The result shows that encoding based on the N-gram method and using Mapping Function to match music melody is feasible.
40
"Automatic caption generation for content-based image information retrieval." 1999. http://library.cuhk.edu.hk/record=b5890055.
Full textAbstract:
Ma, Ka Ho.Thesis (M.Phil.)--Chinese University of Hong Kong, 1999.
Includes bibliographical references (leaves 82-87).
Abstract and appendix in English and Chinese.
Chapter 1 --- Introduction --- p.1
Chapter 1.1 --- Objective of This Research --- p.4
Chapter 1.2 --- Organization of This Thesis --- p.5
Chapter 2 --- Background --- p.6
Chapter 2.1 --- Textual - Image Query Approach --- p.7
Chapter 2.1.1 --- Yahoo! Image Surfer --- p.7
Chapter 2.1.2 --- QBIC (Query By Image Content) --- p.8
Chapter 2.2 --- Feature-based Approach --- p.9
Chapter 2.2.1 --- Texture Thesaurus for Aerial Photos --- p.9
Chapter 2.3 --- Caption-aided Approach --- p.10
Chapter 2.3.1 --- PICTION (Picture and capTION) --- p.10
Chapter 2.3.2 --- MARIE --- p.11
Chapter 2.4 --- Summary --- p.11
Chapter 3 --- Caption Generation --- p.13
Chapter 3.1 --- System Architecture --- p.13
Chapter 3.2 --- Domain Pool --- p.15
Chapter 3.3 --- Image Feature Extraction --- p.16
Chapter 3.3.1 --- Preprocessing --- p.16
Chapter 3.3.2 --- Image Segmentation --- p.17
Chapter 3.4 --- Classification --- p.24
Chapter 3.4.1 --- Self-Organizing Map (SOM) --- p.26
Chapter 3.4.2 --- Learning Vector Quantization (LVQ) --- p.28
Chapter 3.4.3 --- Output of the Classification --- p.30
Chapter 3.5 --- Caption Generation --- p.30
Chapter 3.5.1 --- Phase One: Logical Form Generation --- p.31
Chapter 3.5.2 --- Phase Two: Simplification --- p.32
Chapter 3.5.3 --- Phase Three: Captioning --- p.33
Chapter 3.6 --- Summary --- p.35
Chapter 4 --- Query Examples --- p.37
Chapter 4.1 --- Query Types --- p.37
Chapter 4.1.1 --- Non-content-based Retrieval --- p.38
Chapter 4.1.2 --- Content-based Retrieval --- p.38
Chapter 4.2 --- Hierarchy Graph --- p.41
Chapter 4.3 --- Matching --- p.42
Chapter 4.4 --- Summary --- p.48
Chapter 5 --- Evaluation --- p.49
Chapter 5.1 --- Experimental Set-up --- p.50
Chapter 5.2 --- Experimental Results --- p.51
Chapter 5.2.1 --- Segmentation --- p.51
Chapter 5.2.2 --- Classification --- p.53
Chapter 5.2.3 --- Captioning --- p.55
Chapter 5.2.4 --- Overall Performance --- p.56
Chapter 5.3 --- Observations --- p.57
Chapter 5.4 --- Summary --- p.58
Chapter 6 --- Another Application --- p.59
Chapter 6.1 --- Police Force Crimes Investigation --- p.59
Chapter 6.1.1 --- Image Feature Extraction --- p.61
Chapter 6.1.2 --- Caption Generation --- p.64
Chapter 6.1.3 --- Query --- p.66
Chapter 6.2 --- An Illustrative Example --- p.68
Chapter 6.3 --- Summary --- p.72
Chapter 7 --- Conclusions --- p.74
Chapter 7.1 --- Contribution --- p.77
Chapter 7.2 --- Future Work --- p.78
Bibliography --- p.81
Appendices --- p.88
Chapter A --- Segmentation Result Under Different Parametes --- p.89
Chapter B --- Segmentation Time of 10 Randomly Selected Images --- p.90
Chapter C --- Sample Captions --- p.93
41
Chen, Edwardson, and 陳致生. "Content-based Automatic Annotation and Preference Learning for Music Information Retrieval." Thesis, 2011. http://ndltd.ncl.edu.tw/handle/58uttc.
Full textAbstract:
博士國立清華大學
資訊工程學系
99
Music information retrieval received more and more attention in the past decades. The goal is to find songs, artists, or albums of users’ interests. In this thesis, we focus on two major retrieval approaches, automatic annotation and preference learning recommendation systems. Rather than adopting query-by-example techniques (QBE), searching audio files by a set of semantic concept words is much more natural to associate with music. Such an approach, called query-by-semantic-description (QBSD), needs an accurate and automatic way to help people with tagging lots of audio files. To achieve this demand, we propose an automatic annotation system that uses anti-words for each annotation word based on the concept of supervised multi-class labeling (SML). More specifically, words that are highly associated with the opposite semantic meaning of a word constitute its anti-word set. By modeling both a word and its anti-word set, our annotation system can achieve higher mean per-word precision and recall than the original SML model. Moreover, by constructing the models of the anti-word explicitly, the performance is also significantly improved for the retrieval system. Another major approach for people to discover music is through recommendation which exists frequently in our daily life. Recommenders, such as Amazon, TiVo, and Netflix, adopt collaborative filtering (CF) which often suffers from the so called cold-start problem. However, content-based approach can alleviate this problem since it relies on audio contents instead of users’ past transactions. In the second part of this thesis, we propose a content-based artist recommendation system that can well-predict a user’s tastes. In particular, an artist is characterized by the corresponding acoustical model which is adapted from a universal background model (UBM) through maximum a posterior (MAP) adaptation. These acoustical features, together with their preference rankings, are then used for an ordinal regression algorithm that tries to find a ranking rule which can predict the rank of a new instance. Moreover, an order preserving projection (OPP) algorithm is proposed which is shown to have comparable results with an ordinal regression algorithm, PRank. The proposed linear OPP can also be kernelized to learn the potential nonlinear relationship between music contents and users’ artist rank orders. By introducing the kernel method, we can also efficiently fuse acoustical and symbolic features, i.e. annotation words, under the proposed framework. Experimental results show that the system can successfully predict the user’s tastes and achieve better performance whether using non-linear algorithms of OPP or fusing acoustical and symbolic features.
42
Lin, Chih-Han, and 林志翰. "The Content-based Music Information Retrieval System Based on Music Genre and Emotion." Thesis, 2012. http://ndltd.ncl.edu.tw/handle/49027017243123794873.
Full textAbstract:
碩士國立交通大學
工學院聲音與音樂創意科技碩士學位學程
100
Due to the development of the Internet and smartphone, online music market becomes more and more popular. There’s over millions digital music on the Internet. In order to organize the huge amount of music, an efficient and intelligent music information retrieval system (MIR) is a way to solve this problem. In this thesis, we focus on the content-based music information retrieval system and analyze each music genre and emotion. More specifically, we study some audio feature sets for five difference music characters (Loudness, Tonality, Pitch, Timbre and Rhythm) and music emotion models. The music genre classification approaches described are based on three difference statistical pattern recognition classifiers (SVM, k-NN and LDA). To building the music emotion model are based on SVR. In the result, we implemented a prototype music information retrieval system – MuZhi, which integrate our works of trained music models.
43
"Content-based image retrieval-- a small sample learning approach." 2004. http://library.cuhk.edu.hk/record=b5891962.
Full textAbstract:
Tao Dacheng.Thesis (M.Phil.)--Chinese University of Hong Kong, 2004.
Includes bibliographical references (leaves 70-75).
Abstracts in English and Chinese.
Chapter Chapter 1 --- Introduction --- p.1
Chapter 1.1 --- Content-based Image Retrieval --- p.1
Chapter 1.2 --- SVM based RF in CBIR --- p.3
Chapter 1.3 --- DA based RF in CBIR --- p.4
Chapter 1.4 --- Existing CBIR Engines --- p.5
Chapter 1.5 --- Practical Applications of CBIR --- p.10
Chapter 1.6 --- Organization of this thesis --- p.11
Chapter Chapter 2 --- Statistical Learning Theory and Support Vector Machine --- p.12
Chapter 2.1 --- The Recognition Problem --- p.12
Chapter 2.2 --- Regularization --- p.14
Chapter 2.3 --- The VC Dimension --- p.14
Chapter 2.4 --- Structure Risk Minimization --- p.15
Chapter 2.5 --- Support Vector Machine --- p.15
Chapter 2.6 --- Kernel Space --- p.17
Chapter Chapter 3 --- Discriminant Analysis --- p.18
Chapter 3.1 --- PCA --- p.18
Chapter 3.2 --- KPCA --- p.18
Chapter 3.3 --- LDA --- p.20
Chapter 3.4 --- BDA --- p.20
Chapter 3.5 --- KBDA --- p.21
Chapter Chapter 4 --- Random Sampling Based SVM --- p.24
Chapter 4.1 --- Asymmetric Bagging SVM --- p.25
Chapter 4.2 --- Random Subspace Method SVM --- p.26
Chapter 4.3 --- Asymmetric Bagging RSM SVM --- p.26
Chapter 4.4 --- Aggregation Model --- p.30
Chapter 4.5 --- Dissimilarity Measure --- p.31
Chapter 4.6 --- Computational Complexity Analysis --- p.31
Chapter 4.7 --- QueryGo Image Retrieval System --- p.32
Chapter 4.8 --- Toy Experiments --- p.35
Chapter 4.9 --- Statistical Experimental Results --- p.36
Chapter Chapter 5 --- SSS Problems in KBDA RF --- p.42
Chapter 5.1 --- DKBDA --- p.43
Chapter 5.1.1 --- DLDA --- p.43
Chapter 5.1.2 --- DKBDA --- p.43
Chapter 5.2 --- NKBDA --- p.48
Chapter 5.2.1 --- NLDA --- p.48
Chapter 5.2.2 --- NKBDA --- p.48
Chapter 5.3 --- FKBDA --- p.49
Chapter 5.3.1 --- FLDA --- p.49
Chapter 5.3.2 --- FKBDA --- p.49
Chapter 5.4 --- Experimental Results --- p.50
Chapter Chapter 6 --- NDA based RF for CBIR --- p.52
Chapter 6.1 --- NDA --- p.52
Chapter 6.2 --- SSS Problem in NDA --- p.53
Chapter 6.2.1 --- Regularization method --- p.53
Chapter 6.2.2 --- Null-space method --- p.54
Chapter 6.2.3 --- Full-space method --- p.54
Chapter 6.3 --- Experimental results --- p.55
Chapter 6.3.1 --- K nearest neighbor evaluation for NDA --- p.55
Chapter 6.3.2 --- SSS problem --- p.56
Chapter 6.3.3 --- Evaluation experiments --- p.57
Chapter Chapter 7 --- Medical Image Classification --- p.59
Chapter 7.1 --- Introduction --- p.59
Chapter 7.2 --- Region-based Co-occurrence Matrix Texture Feature --- p.60
Chapter 7.3 --- Multi-level Feature Selection --- p.62
Chapter 7.4 --- Experimental Results --- p.63
Chapter 7.4.1 --- Data Set --- p.64
Chapter 7.4.2 --- Classification Using Traditional Features --- p.65
Chapter 7.4.3 --- Classification Using the New Features --- p.66
Chapter Chapter 8 --- Conclusion --- p.68
Bibliography --- p.70
44
Sun, Donghu Liu Xiuwen. "A study of image representations for content-based image retrieval." 2004. http://etd.lib.fsu.edu/theses/available/etd-04122004-155732.
Full textAbstract:
Thesis (M.S.)--Florida State University, 2004.Advisor: Dr. Xiuwen Liu, Florida State University, College of Arts and Sciences, Dept. of Computer Science. Title and description from dissertation home page (viewed June 15, 2004). Includes bibliographical references.
45
"Learning on relevance feedback in content-based image retrieval." 2004. http://library.cuhk.edu.hk/record=b5892070.
Full textAbstract:
Hoi, Chu-Hong.Thesis (M.Phil.)--Chinese University of Hong Kong, 2004.
Includes bibliographical references (leaves 89-103).
Abstracts in English and Chinese.
Abstract --- p.i
Acknowledgement --- p.iv
Chapter 1 --- Introduction --- p.1
Chapter 1.1 --- Content-based Image Retrieval --- p.1
Chapter 1.2 --- Relevance Feedback --- p.3
Chapter 1.3 --- Contributions --- p.4
Chapter 1.4 --- Organization of This Work --- p.6
Chapter 2 --- Background --- p.8
Chapter 2.1 --- Relevance Feedback --- p.8
Chapter 2.1.1 --- Heuristic Weighting Methods --- p.9
Chapter 2.1.2 --- Optimization Formulations --- p.10
Chapter 2.1.3 --- Various Machine Learning Techniques --- p.11
Chapter 2.2 --- Support Vector Machines --- p.12
Chapter 2.2.1 --- Setting of the Learning Problem --- p.12
Chapter 2.2.2 --- Optimal Separating Hyperplane --- p.13
Chapter 2.2.3 --- Soft-Margin Support Vector Machine --- p.15
Chapter 2.2.4 --- One-Class Support Vector Machine --- p.16
Chapter 3 --- Relevance Feedback with Biased SVM --- p.18
Chapter 3.1 --- Introduction --- p.18
Chapter 3.2 --- Biased Support Vector Machine --- p.19
Chapter 3.3 --- Relevance Feedback Using Biased SVM --- p.22
Chapter 3.3.1 --- Advantages of BSVM in Relevance Feedback --- p.22
Chapter 3.3.2 --- Relevance Feedback Algorithm by BSVM --- p.23
Chapter 3.4 --- Experiments --- p.24
Chapter 3.4.1 --- Datasets --- p.24
Chapter 3.4.2 --- Image Representation --- p.25
Chapter 3.4.3 --- Experimental Results --- p.26
Chapter 3.5 --- Discussions --- p.29
Chapter 3.6 --- Summary --- p.30
Chapter 4 --- Optimizing Learning with SVM Constraint --- p.31
Chapter 4.1 --- Introduction --- p.31
Chapter 4.2 --- Related Work and Motivation --- p.33
Chapter 4.3 --- Optimizing Learning with SVM Constraint --- p.35
Chapter 4.3.1 --- Problem Formulation and Notations --- p.35
Chapter 4.3.2 --- Learning boundaries with SVM --- p.35
Chapter 4.3.3 --- OPL for the Optimal Distance Function --- p.38
Chapter 4.3.4 --- Overall Similarity Measure with OPL and SVM --- p.40
Chapter 4.4 --- Experiments --- p.41
Chapter 4.4.1 --- Datasets --- p.41
Chapter 4.4.2 --- Image Representation --- p.42
Chapter 4.4.3 --- Performance Evaluation --- p.43
Chapter 4.4.4 --- Complexity and Time Cost Evaluation --- p.45
Chapter 4.5 --- Discussions --- p.47
Chapter 4.6 --- Summary --- p.48
Chapter 5 --- Group-based Relevance Feedback --- p.49
Chapter 5.1 --- Introduction --- p.49
Chapter 5.2 --- SVM Ensembles --- p.50
Chapter 5.3 --- Group-based Relevance Feedback Using SVM Ensembles --- p.51
Chapter 5.3.1 --- (x+l)-class Assumption --- p.51
Chapter 5.3.2 --- Proposed Architecture --- p.52
Chapter 5.3.3 --- Strategy for SVM Combination and Group Ag- gregation --- p.52
Chapter 5.4 --- Experiments --- p.54
Chapter 5.4.1 --- Experimental Implementation --- p.54
Chapter 5.4.2 --- Performance Evaluation --- p.55
Chapter 5.5 --- Discussions --- p.56
Chapter 5.6 --- Summary --- p.57
Chapter 6 --- Log-based Relevance Feedback --- p.58
Chapter 6.1 --- Introduction --- p.58
Chapter 6.2 --- Related Work and Motivation --- p.60
Chapter 6.3 --- Log-based Relevance Feedback Using SLSVM --- p.61
Chapter 6.3.1 --- Problem Statement --- p.61
Chapter 6.3.2 --- Soft Label Support Vector Machine --- p.62
Chapter 6.3.3 --- LRF Algorithm by SLSVM --- p.64
Chapter 6.4 --- Experimental Results --- p.66
Chapter 6.4.1 --- Datasets --- p.66
Chapter 6.4.2 --- Image Representation --- p.66
Chapter 6.4.3 --- Experimental Setup --- p.67
Chapter 6.4.4 --- Performance Comparison --- p.68
Chapter 6.5 --- Discussions --- p.73
Chapter 6.6 --- Summary --- p.75
Chapter 7 --- Application: Web Image Learning --- p.76
Chapter 7.1 --- Introduction --- p.76
Chapter 7.2 --- A Learning Scheme for Searching Semantic Concepts --- p.77
Chapter 7.2.1 --- Searching and Clustering Web Images --- p.78
Chapter 7.2.2 --- Learning Semantic Concepts with Relevance Feed- back --- p.73
Chapter 7.3 --- Experimental Results --- p.79
Chapter 7.3.1 --- Dataset and Features --- p.79
Chapter 7.3.2 --- Performance Evaluation --- p.80
Chapter 7.4 --- Discussions --- p.82
Chapter 7.5 --- Summary --- p.82
Chapter 8 --- Conclusions and Future Work --- p.84
Chapter 8.1 --- Conclusions --- p.84
Chapter 8.2 --- Future Work --- p.85
Chapter A --- List of Publications --- p.87
Bibliography --- p.103
46
"Biased classification for relevance feedback in content-based image retrieval." 2007. http://library.cuhk.edu.hk/record=b5893182.
Full textAbstract:
Peng, Xiang.Thesis (M.Phil.)--Chinese University of Hong Kong, 2007.
Includes bibliographical references (leaves 98-115).
Abstracts in English and Chinese.
Abstract --- p.i
Acknowledgement --- p.iv
Chapter 1 --- Introduction --- p.1
Chapter 1.1 --- Problem Statement --- p.3
Chapter 1.2 --- Major Contributions --- p.6
Chapter 1.3 --- Thesis Outline --- p.7
Chapter 2 --- Background Study --- p.9
Chapter 2.1 --- Content-based Image Retrieval --- p.9
Chapter 2.1.1 --- Image Representation --- p.11
Chapter 2.1.2 --- High Dimensional Indexing --- p.15
Chapter 2.1.3 --- Image Retrieval Systems Design --- p.16
Chapter 2.2 --- Relevance Feedback --- p.19
Chapter 2.2.1 --- Self-Organizing Map in Relevance Feedback --- p.20
Chapter 2.2.2 --- Decision Tree in Relevance Feedback --- p.22
Chapter 2.2.3 --- Bayesian Classifier in Relevance Feedback --- p.24
Chapter 2.2.4 --- Nearest Neighbor Search in Relevance Feedback --- p.25
Chapter 2.2.5 --- Support Vector Machines in Relevance Feedback --- p.26
Chapter 2.3 --- Imbalanced Classification --- p.29
Chapter 2.4 --- Active Learning --- p.31
Chapter 2.4.1 --- Uncertainly-based Sampling --- p.33
Chapter 2.4.2 --- Error Reduction --- p.34
Chapter 2.4.3 --- Batch Selection --- p.35
Chapter 2.5 --- Convex Optimization --- p.35
Chapter 2.5.1 --- Overview of Convex Optimization --- p.35
Chapter 2.5.2 --- Linear Program --- p.37
Chapter 2.5.3 --- Quadratic Program --- p.37
Chapter 2.5.4 --- Quadratically Constrained Quadratic Program --- p.37
Chapter 2.5.5 --- Cone Program --- p.38
Chapter 2.5.6 --- Semi-definite Program --- p.39
Chapter 3 --- Imbalanced Learning with BMPM for CBIR --- p.40
Chapter 3.1 --- Research Motivation --- p.41
Chapter 3.2 --- Background Review --- p.42
Chapter 3.2.1 --- Relevance Feedback for CBIR --- p.42
Chapter 3.2.2 --- Minimax Probability Machine --- p.42
Chapter 3.2.3 --- Extensions of Minimax Probability Machine --- p.44
Chapter 3.3 --- Relevance Feedback using BMPM --- p.45
Chapter 3.3.1 --- Model Definition --- p.45
Chapter 3.3.2 --- Advantages of BMPM in Relevance Feedback --- p.46
Chapter 3.3.3 --- Relevance Feedback Framework by BMPM --- p.47
Chapter 3.4 --- Experimental Results --- p.47
Chapter 3.4.1 --- Experiment Datasets --- p.48
Chapter 3.4.2 --- Performance Evaluation --- p.50
Chapter 3.4.3 --- Discussions --- p.53
Chapter 3.5 --- Summary --- p.53
Chapter 4 --- BMPM Active Learning for CBIR --- p.55
Chapter 4.1 --- Problem Statement and Motivation --- p.55
Chapter 4.2 --- Background Review --- p.57
Chapter 4.3 --- Relevance Feedback by BMPM Active Learning . --- p.58
Chapter 4.3.1 --- Active Learning Concept --- p.58
Chapter 4.3.2 --- General Approaches for Active Learning . --- p.59
Chapter 4.3.3 --- Biased Minimax Probability Machine --- p.60
Chapter 4.3.4 --- Proposed Framework --- p.61
Chapter 4.4 --- Experimental Results --- p.63
Chapter 4.4.1 --- Experiment Setup --- p.64
Chapter 4.4.2 --- Performance Evaluation --- p.66
Chapter 4.5 --- Summary --- p.68
Chapter 5 --- Large Scale Learning with BMPM --- p.70
Chapter 5.1 --- Introduction --- p.71
Chapter 5.1.1 --- Motivation --- p.71
Chapter 5.1.2 --- Contribution --- p.72
Chapter 5.2 --- Background Review --- p.72
Chapter 5.2.1 --- Second Order Cone Program --- p.72
Chapter 5.2.2 --- General Methods for Large Scale Problems --- p.73
Chapter 5.2.3 --- Biased Minimax Probability Machine --- p.75
Chapter 5.3 --- Efficient BMPM Training --- p.78
Chapter 5.3.1 --- Proposed Strategy --- p.78
Chapter 5.3.2 --- Kernelized BMPM and Its Solution --- p.81
Chapter 5.4 --- Experimental Results --- p.82
Chapter 5.4.1 --- Experimental Testbeds --- p.83
Chapter 5.4.2 --- Experimental Settings --- p.85
Chapter 5.4.3 --- Performance Evaluation --- p.87
Chapter 5.5 --- Summary --- p.92
Chapter 6 --- Conclusion and Future Work --- p.93
Chapter 6.1 --- Conclusion --- p.93
Chapter 6.2 --- Future Work --- p.94
Chapter A --- List of Symbols and Notations --- p.96
Chapter B --- List of Publications --- p.98
Bibliography --- p.100
47
張厥煒. "= Structured video computing and content-based retrieval in a video information system." Thesis, 1996. http://ndltd.ncl.edu.tw/handle/65099314673783857451.
Full text
48
"Rival penalized competitive learning for content-based indexing." 1998. http://library.cuhk.edu.hk/record=b5889612.
Full textAbstract:
by Lau Tak Kan.Thesis (M.Phil.)--Chinese University of Hong Kong, 1998.
Includes bibliographical references (leaves 100-108).
Abstract also in Chinese.
Chapter 1 --- Introduction --- p.1
Chapter 1.1 --- Background --- p.1
Chapter 1.2 --- Problem Defined --- p.5
Chapter 1.3 --- Contributions --- p.5
Chapter 1.4 --- Thesis Organization --- p.7
Chapter 2 --- Content-based Retrieval Multimedia Database Background and Indexing Problem --- p.8
Chapter 2.1 --- Feature Extraction --- p.8
Chapter 2.2 --- Nearest-neighbor Search --- p.10
Chapter 2.3 --- Content-based Indexing Methods --- p.15
Chapter 2.4 --- Indexing Problem --- p.22
Chapter 3 --- Data Clustering Methods for Indexing --- p.25
Chapter 3.1 --- Proposed Solution to Indexing Problem --- p.25
Chapter 3.2 --- Brief Description of Several Clustering Methods --- p.26
Chapter 3.2.1 --- K-means --- p.26
Chapter 3.2.2 --- Competitive Learning (CL) --- p.27
Chapter 3.2.3 --- Rival Penalized Competitive Learning (RPCL) --- p.29
Chapter 3.2.4 --- General Hierarchical Clustering Methods --- p.31
Chapter 3.3 --- Why RPCL? --- p.32
Chapter 4 --- Non-hierarchical RPCL Indexing --- p.33
Chapter 4.1 --- The Non-hierarchical Approach --- p.33
Chapter 4.2 --- Performance Experiments --- p.34
Chapter 4.2.1 --- Experimental Setup --- p.35
Chapter 4.2.2 --- Experiment 1: Test for Recall and Precision Performance --- p.38
Chapter 4.2.3 --- Experiment 2: Test for Different Sizes of Input Data Sets --- p.45
Chapter 4.2.4 --- Experiment 3: Test for Different Numbers of Dimensions --- p.49
Chapter 4.2.5 --- Experiment 4: Compare with Actual Nearest-neighbor Results --- p.53
Chapter 4.3 --- Chapter Summary --- p.55
Chapter 5 --- Hierarchical RPCL Indexing --- p.56
Chapter 5.1 --- The Hierarchical Approach --- p.56
Chapter 5.2 --- The Hierarchical RPCL Binary Tree (RPCL-b-tree) --- p.58
Chapter 5.3 --- Insertion --- p.61
Chapter 5.4 --- Deletion --- p.63
Chapter 5.5 --- Searching --- p.63
Chapter 5.6 --- Experiments --- p.69
Chapter 5.6.1 --- Experimental Setup --- p.69
Chapter 5.6.2 --- Experiment 5: Test for Different Node Sizes --- p.72
Chapter 5.6.3 --- Experiment 6: Test for Different Sizes of Data Sets --- p.75
Chapter 5.6.4 --- Experiment 7: Test for Different Data Distributions --- p.78
Chapter 5.6.5 --- Experiment 8: Test for Different Numbers of Dimensions --- p.80
Chapter 5.6.6 --- Experiment 9: Test for Different Numbers of Database Ob- jects Retrieved --- p.83
Chapter 5.6.7 --- Experiment 10: Test with VP-tree --- p.86
Chapter 5.7 --- Discussion --- p.90
Chapter 5.8 --- A Relationship Formula --- p.93
Chapter 5.9 --- Chapter Summary --- p.96
Chapter 6 --- Conclusion --- p.97
Chapter 6.1 --- Future Works --- p.97
Chapter 6.2 --- Conclusion --- p.98
Bibliography --- p.100
49
"Design, implementation, and evaluation of scalable content-based image retrieval techniques." 2007. http://library.cuhk.edu.hk/record=b5893292.
Full textAbstract:
Wong, Yuk Man.Thesis (M.Phil.)--Chinese University of Hong Kong, 2007.
Includes bibliographical references (leaves 95-100).
Abstracts in English and Chinese.
Abstract --- p.ii
Acknowledgement --- p.v
Chapter 1 --- Introduction --- p.1
Chapter 1.1 --- Overview --- p.1
Chapter 1.2 --- Contribution --- p.3
Chapter 1.3 --- Organization of This Work --- p.5
Chapter 2 --- Literature Review --- p.6
Chapter 2.1 --- Content-based Image Retrieval --- p.6
Chapter 2.1.1 --- Query Technique --- p.6
Chapter 2.1.2 --- Relevance Feedback --- p.7
Chapter 2.1.3 --- Previously Proposed CBIR systems --- p.7
Chapter 2.2 --- Invariant Local Feature --- p.8
Chapter 2.3 --- Invariant Local Feature Detector --- p.9
Chapter 2.3.1 --- Harris Corner Detector --- p.9
Chapter 2.3.2 --- DOG Extrema Detector --- p.10
Chapter 2.3.3 --- Harris-Laplacian Corner Detector --- p.13
Chapter 2.3.4 --- Harris-Affine Covariant Detector --- p.14
Chapter 2.4 --- Invariant Local Feature Descriptor --- p.15
Chapter 2.4.1 --- Scale Invariant Feature Transform (SIFT) --- p.15
Chapter 2.4.2 --- Shape Context --- p.17
Chapter 2.4.3 --- PCA-SIFT --- p.18
Chapter 2.4.4 --- Gradient Location and Orientation Histogram (GLOH) --- p.19
Chapter 2.4.5 --- Geodesic-Intensity Histogram (GIH) --- p.19
Chapter 2.4.6 --- Experiment --- p.21
Chapter 2.5 --- Feature Matching --- p.27
Chapter 2.5.1 --- Matching Criteria --- p.27
Chapter 2.5.2 --- Distance Measures --- p.28
Chapter 2.5.3 --- Searching Techniques --- p.29
Chapter 3 --- A Distributed Scheme for Large-Scale CBIR --- p.31
Chapter 3.1 --- Overview --- p.31
Chapter 3.2 --- Related Work --- p.33
Chapter 3.3 --- Scalable Content-Based Image Retrieval Scheme --- p.34
Chapter 3.3.1 --- Overview of Our Solution --- p.34
Chapter 3.3.2 --- Locality-Sensitive Hashing --- p.34
Chapter 3.3.3 --- Scalable Indexing Solutions --- p.35
Chapter 3.3.4 --- Disk-Based Multi-Partition Indexing --- p.36
Chapter 3.3.5 --- Parallel Multi-Partition Indexing --- p.37
Chapter 3.4 --- Feature Representation --- p.43
Chapter 3.5 --- Empirical Evaluation --- p.44
Chapter 3.5.1 --- Experimental Testbed --- p.44
Chapter 3.5.2 --- Performance Evaluation Metrics --- p.44
Chapter 3.5.3 --- Experimental Setup --- p.45
Chapter 3.5.4 --- Experiment I: Disk-Based Multi-Partition Indexing Approach --- p.45
Chapter 3.5.5 --- Experiment II: Parallel-Based Multi-Partition Indexing Approach --- p.48
Chapter 3.6 --- Application to WWW Image Retrieval --- p.55
Chapter 3.7 --- Summary --- p.55
Chapter 4 --- Image Retrieval System for IND Detection --- p.60
Chapter 4.1 --- Overview --- p.60
Chapter 4.1.1 --- Motivation --- p.60
Chapter 4.1.2 --- Related Work --- p.61
Chapter 4.1.3 --- Objective --- p.62
Chapter 4.1.4 --- Contribution --- p.63
Chapter 4.2 --- Database Construction --- p.63
Chapter 4.2.1 --- Image Representations --- p.63
Chapter 4.2.2 --- Index Construction --- p.64
Chapter 4.2.3 --- Keypoint and Image Lookup Tables --- p.67
Chapter 4.3 --- Database Query --- p.67
Chapter 4.3.1 --- Matching Strategies --- p.68
Chapter 4.3.2 --- Verification Processes --- p.71
Chapter 4.3.3 --- Image Voting --- p.75
Chapter 4.4 --- Performance Evaluation --- p.76
Chapter 4.4.1 --- Evaluation Metrics --- p.76
Chapter 4.4.2 --- Results --- p.77
Chapter 4.4.3 --- Summary --- p.81
Chapter 5 --- Shape-SIFT Feature Descriptor --- p.82
Chapter 5.1 --- Overview --- p.82
Chapter 5.2 --- Related Work --- p.83
Chapter 5.3 --- SHAPE-SIFT Descriptors --- p.84
Chapter 5.3.1 --- Orientation assignment --- p.84
Chapter 5.3.2 --- Canonical orientation determination --- p.84
Chapter 5.3.3 --- Keypoint descriptor --- p.87
Chapter 5.4 --- Performance Evaluation --- p.88
Chapter 5.5 --- Summary --- p.90
Chapter 6 --- Conclusions and Future Work --- p.92
Chapter 6.1 --- Conclusions --- p.92
Chapter 6.2 --- Future Work --- p.93
Chapter A --- Publication --- p.94
Bibliography --- p.95
50
"Redundancy on content-based indexing." 1997. http://library.cuhk.edu.hk/record=b5889125.
Full textAbstract:
by Cheung King Lum Kingly.Thesis (M.Phil.)--Chinese University of Hong Kong, 1997.
Includes bibliographical references (leaves 108-110).
Abstract --- p.ii
Acknowledgement --- p.iii
Chapter 1 --- Introduction --- p.1
Chapter 1.1 --- Motivation --- p.1
Chapter 1.2 --- Problems in Content-Based Indexing --- p.2
Chapter 1.3 --- Contributions --- p.3
Chapter 1.4 --- Thesis Organization --- p.4
Chapter 2 --- Content-Based Indexing Structures --- p.5
Chapter 2.1 --- R-Tree --- p.6
Chapter 2.2 --- R+-Tree --- p.8
Chapter 2.3 --- R*-Tree --- p.11
Chapter 3 --- Searching in Both R-Tree and R*-Tree --- p.15
Chapter 3.1 --- Exact Search --- p.15
Chapter 3.2 --- Nearest Neighbor Search --- p.19
Chapter 3.2.1 --- Definition of Searching Metrics --- p.19
Chapter 3.2.2 --- Pruning Heuristics --- p.21
Chapter 3.2.3 --- Nearest Neighbor Search Algorithm --- p.24
Chapter 3.2.4 --- Generalization to N-Nearest Neighbor Search --- p.25
Chapter 4 --- An Improved Nearest Neighbor Search Algorithm for R-Tree --- p.29
Chapter 4.1 --- Introduction --- p.29
Chapter 4.2 --- New Pruning Heuristics --- p.31
Chapter 4.3 --- An Improved Nearest Neighbor Search Algorithm --- p.34
Chapter 4.4 --- Replacing Heuristics --- p.36
Chapter 4.5 --- N-Nearest Neighbor Search --- p.41
Chapter 4.6 --- Performance Evaluation --- p.45
Chapter 5 --- Overlapping Nodes in R-Tree and R*-Tree --- p.53
Chapter 5.1 --- Overlapping Nodes --- p.54
Chapter 5.2 --- Problem Induced By Overlapping Nodes --- p.57
Chapter 5.2.1 --- Backtracking --- p.57
Chapter 5.2.2 --- Inefficient Exact Search --- p.57
Chapter 5.2.3 --- Inefficient Nearest Neighbor Search --- p.60
Chapter 6 --- Redundancy On R-Tree --- p.64
Chapter 6.1 --- Motivation --- p.64
Chapter 6.2 --- Adding Redundancy on Index Tree --- p.65
Chapter 6.3 --- R-Tree with Redundancy --- p.66
Chapter 6.3.1 --- Previous Models of R-Tree with Redundancy --- p.66
Chapter 6.3.2 --- Redundant R-Tree --- p.70
Chapter 6.3.3 --- Level List --- p.71
Chapter 6.3.4 --- Inserting Redundancy to R-Tree --- p.72
Chapter 6.3.5 --- Properties of Redundant R-Tree --- p.77
Chapter 7 --- Searching in Redundant R-Tree --- p.82
Chapter 7.1 --- Exact Search --- p.82
Chapter 7.2 --- Nearest Neighbor Search --- p.86
Chapter 7.3 --- Avoidance of Multiple Accesses --- p.89
Chapter 8 --- Experiment --- p.90
Chapter 8.1 --- Experimental Setup --- p.90
Chapter 8.2 --- Exact Search --- p.91
Chapter 8.2.1 --- Clustered Data --- p.91
Chapter 8.2.2 --- Real Data --- p.93
Chapter 8.3 --- Nearest Neighbor Search --- p.95
Chapter 8.3.1 --- Clustered Data --- p.95
Chapter 8.3.2 --- Uniform Data --- p.98
Chapter 8.3.3 --- Real Data --- p.100
Chapter 8.4 --- Discussion --- p.102
Chapter 9 --- Conclusions and Future Research --- p.105
Chapter 9.1 --- Conclusions --- p.105
Chapter 9.2 --- Future Research --- p.106
Bibliography --- p.108