Dissertations / Theses on the topic 'Video recordings – Computer programs'

Thesis: S.M., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2018.
Cataloged from PDF version of thesis.
Includes bibliographical references (pages 87-92).
Accurate quantification of neurodegenerative disease progression and neurocognitive decline is an ongoing challenge that muddles efforts to understand and treat these conditions. The golden standard to track neurodegenerative disease progression is a neurocognitive assessment administered through a series of forms and questionnaires, both of which suffer from high retest variability. Thus, there is a need for a validated biomarker that can accurately track neurocognitive decline or disease progression unobtrusively. Clinical studies have shown that saccade latency - a measure of eye movement reaction time - can be significantly different between healthy subjects and neurodegenerative disease patients. Hence, we propose and test a system that enables measurement and tracking of saccade latency outside of the clinical environment, without relying on supplemental or special-purpose illumination. iTracker-face, a variant of a deep-learning approach that estimates gaze on mobile devices, was tailored to measure saccade latency on video sequences obtained with an iPhone 6. Additionally, our model-based approach for saccade latency estimation provides a means to automatically detect eye position traces that should not be included in further analysis (AUC = 0.92). We recorded over 20,000 saccade latencies in 30 healthy subjects and observed that there is a substantial amount of intra- and inter- subject variability in the mean saccade latency ([mu]) and associated standard deviation (sigma). Among the subjects with the most saccade latency measurements, [mu] ranged from 128 to 173 ms, while a varied from 24 to 53 ms. Our results suggest that accurate and robust saccade latency determination is feasible using consumer-grade cameras and might therefore enable unobtrusive tracking of neurodegenerative disease progression.
by Gladynel Saavedra-Peña.
S.M.

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 February 20, 2008) Vita. Includes bibliographical references.

This dissertation presents and evaluates a method to the video indexing problem by investigating a categorization method that transcribes audio content through Automatic Speech Recognition (ASR) combined with Dynamic Contextualization (DC), Phonetic Disparity Search (PDS) and Metaphone indexation. The suggested approach applies genome pattern matching algorithms with computational summarization to build a database infrastructure that provides an indexed summary of the original audio content. PDS complements the contextual phoneme indexing approach by optimizing topic seek performance and accuracy in large video content structures. A prototype was established to translate news broadcast video into text and phonemes automatically by using ASR utterance conversions. Each phonetic utterance extraction was then categorized, converted to Metaphones, and stored in a repository with contextual topical information attached and indexed for posterior search analysis. Following the original design strategy, a custom parallel interface was built to measure the capabilities of dissimilar phonetic queries and provide an interface for result analysis. The postulated solution provides evidence of a superior topic matching when compared to traditional word and phoneme search methods. Experimental results demonstrate that PDS can be 3.7% better than the same phoneme query, Metaphone search proved to be 154.6% better than the same phoneme seek and 68.1 % better than the equivalent word search.
ID: 030423400; System requirements: World Wide Web browser and PDF reader.; Mode of access: World Wide Web.; Thesis (Ph.D.)--University of Central Florida, 2010.; Includes bibliographical references (p. 228-236).
Ph.D.
Doctorate
Department of Electrical Engineering and Computer Science
Engineering and Computer Science

Thesis (Ph. D.)--Illinois State University, 1985.
Title from title page screen, viewed June 23, 2005. Dissertation Committee: John McCarthy (chair), Harry Campbell, Ronald Halinski, Everett Israel, J.H. McGrath. Includes bibliographical references (leaves 128-134) and abstract. Also available in print.

Thesis (Ed.D.)--Teachers College, Columbia University, 1996.
Typescript; issued also on microfilm. Sponsor: Frank L. Smith. Dissertation Committee: Jeanette E. Fleischer. Includes bibliographical references (leaves 128-130).

International Telemetering Conference Proceedings / October 25-28, 1993 / Riviera Hotel and Convention Center, Las Vegas, Nevada
In today's world, the importance of training for telemetry systems continues to grow as new technologies provide users with ever-increasing capabilities. Successful training programs ensure telemetry systems quickly become operational yielding the acquisition of critical test data. Over the years, training programs have been greatly impacted by the changes in defense contracts -- specifically funding. Today's aggressive telemetry market requires contractors to develop complex telemetry systems within the constraints of Firm Fixed price (FFP) contracts and within very short schedules. As a result of these conditions, training programs have changed significantly over the last ten years. Projects which used to have dedicated training personnel (instructors, technical writers, etc.) now rely on the system developers to provide the training. In actuality, the quality of training has improved with this new approach. Now students benefit from having the most knowledgeable personnel teach them about the system and, often times, latent problems with the system are efficiently identified and corrected. This paper will summarize the evolution of training programs for telemetry systems developed by Computer Sciences Corporation. The benefits of a scenario where the system architects train the end users with the use of increased "hands-on" training will be explored.

Thesis (MEng)--Stellenbosch University, 2014.
ENGLISH ABSTRACT: We design and implement a prototype digital watermarking scheme that uses spread-transform dither modulation to embed hidden binary message signals into uncompressed digital video signals. The hidden messages are intended to identify the original recipient of a particular copy of a digital video, so that in the event that a pirate copy of that video is found, the recipient can be incriminated as a pirate. We test the ability of the watermark signal to survive various modifications to the watermarked video, including lossy compression, frame deletion, and amplitude scaling. We find that the scheme performs well when watermarked videos remain uncompressed, but the scheme is not sufficiently robust against lossy compression. We also implement a dynamic traitor tracing scheme based on the scheme of Fiat and Tassa, which is designed to identify a group of pirates who combine their watermarked videos in an attempt to defeat the watermarking scheme. We test the traitor tracing scheme in a simulated digitally watermarked video broadcast system where a pirate rebroadcast is monitored. We find that when the pirate rebroadcast is delayed, our traitor tracing scheme identifies and disconnects the pirates more quickly than the scheme of Fiat and Tassa.
AFRIKAANSE OPSOMMING: Ons ontwerp en implementeer ’n digitale watermerkskema-prototipe wat strektransform-bibbermodulasie gebruik om binêre boodskapseine in digitale videoseine te verskuil. Die doel van die versteekte boodskappe is om die oorspronklike ontvanger van ’n bepaalde digitale video-afskrif te identifiseer. Indien ’n onwettige afskrif van die video ontdek word, kan die oorspronklike ontvanger daarvoor verantwoordelik gehou word. Ons toets die vermoë van die watermerksein om verskeie veranderinge aan die gewatermerkte video te oorleef, byvoorbeeld kompressie, die verwydering van videorame, en amplitudeskalering. Die skema vaar goed in die afwesigheid van videokompressie, maar is nie robuus wanneer die video beduidend saamgepers word nie. Ons implementeer ook ’n dinamiese skema vir verraaier-opsporing, gebaseer op ’n tegniek deur Fiat en Tassa. Die doel van hierdie skema is om groepe gebruikers te identifiseer wat saamwerk om ’n watermerkskema te kul deur gewatermerkte videos te kombineer. Ons toets die verraaier-opsporingskema in ’n simulasie van ’n gewatermerkte video-uitsendingstelsel wat ’n roofuitsending monitor. Wanneer die roofuitsending vertraag is, vind ons dat die voorgestelde verraaier-opsporingskema die skuldige gebruikers vinniger uitwys as die skema van Fiat en Tassa.

Combat arms units (both Marine and Army) often do not have enough people, time and resources to properly train collective tasks at the squad level. Resources are often retained by higher headquarters due to tight deployment schedules, land restrictions, logistics constraints and a myriad of other reasons. Due to the current operational demands of combat arms brigades and regiments, the reality of limited resources is often a contributing factor in poor performance at the squad level. Leaders at all levels will need to look for innovative ways to sustain training levels at the small unit level. The scope of this study examined the collective and leader tasks that are required for successful execution of Infantry squad missions (using the Army Training and Evaluation Plan ARTEP 7-8 Drill), and how those tasks could be trained with the use of commercial off-the-shelf multiplayer gaming software. The end-state of this research study is to provide initial analysis on what collective skills games can be used to train at the Infantry squad level, and develop a training model recommendation for the integration of this tool into existing unit plans.

Shadows in indoor scenarios are usually characterized with multiple light sources that produce complex shadow patterns of a single object. Without removing shadow, the foreground object tends to be erroneously segmented. The inconsistent hue and intensity of shadows make automatic removal a challenging task. In this thesis, a dynamic thresholding and transfer learning-based method for removing shadows is proposed. The method suppresses light shadows with a dynamically computed threshold and removes dark shadows using an online learning strategy that is built upon a base classifier trained with manually annotated examples and refined with the automatically identified examples in the new videos. Experimental results demonstrate that despite variation of lighting conditions in videos our proposed method is able to adapt to the videos and remove shadows effectively. The sensitivity of shadow detection changes slightly with different confidence levels used in example selection for classifier retraining and high confidence level usually yields better performance with less retraining iterations.

A significant number of action recognition research efforts use spatio-temporal interest point detectors for feature extraction. Although the extracted features provide useful information for recognizing actions, a significant number of them contain irrelevant motion and background clutter. In many cases, the extracted features are included as is in the classification pipeline, and sophisticated noise removal techniques are subsequently used to alleviate their effect on classification. We introduce a new action database, created from the Weizmann database, that reveals a significant weakness in systems based on popular cuboid descriptors. Experiments show that introducing complex backgrounds, stationary or dynamic, into the video causes a significant degradation in recognition performance. Moreover, this degradation cannot be fixed by fine-tuning the system or selecting better interest points. Instead, we show that the problem lies at the descriptor level and must be addressed by modifying descriptors.
ID: 030423225; System requirements: World Wide Web browser and PDF reader.; Mode of access: World Wide Web.; Thesis (M.S.)--University of Central Florida, 2011.; Includes bibliographical references (p. 40-41).
M.S.
Masters
Electrical Engineering and Computer Science
Engineering and Computer Science

Conselho Nacional do Desenvolvimento Científico e Tecnológico (CNPq)
É implementado em hardware, por meio da linguagem VHDL, um método de avaliação objetiva de qualidade de vídeo digital. Sendo um processo computacionalmente custoso em software, investiga-se sua implementação em hardware. O método implementado, chamado de NRVQA-LM, utiliza seis características espaço-temporais extraídas de diferentes vídeos para chegar a um escore de qualidade. São estudadas essas características e planejada a sua implementação de forma otimizada, a fim de aproveitar as vantagens de plataformas de lógica reconfigurável, como as FPGAS. Durante o desenvolvimento foi necessário o estudo de ferramentas não usuais da linguagem VHDL, tais como as aritméticas de ponto fixo e flutuante e a escrita de funções matemáticas. Os resultados mostram alta correlação com os valores das características e dos escores de qualidade em relação ao método em software. A implementação se mostrou custosa em termos de recursos lógicos, especialmente devido à necessidade de se armazenar um quadro de vídeo inteiro, mas eficiente graças à característica de paralelismo das FPGAs, executando cálculos entre 20 e 40 vezes mais rapidamente que em uma linguagem de alto nível como o MATLAB. A aritmética ponto fixo mostrou-se vantajosa em relação ao ponto flutuante, principalmente no que tange à frequência de operação.
It is implemented in hardware an objective digital video evaluation method, using the VHDL language. As a computationally expensive process in software, it is investigated its implementation in a hardware platform. The implemented method, named NRVQA-LM, employs six spatio-temporal features extracted from different videos in order to obtain a quality score. These features are studied and the implementation is designed to be developed in an optimized way, in order to explore the benefits of reprogammable logic platforms, such as FPGAs. During the development it was necessary to study non-recurrent tools of the VHDL language, such as fixed- and floating-point arithmetics and the writing of math functions. Results shows high correlation between the calculated scores of the hardware and the original software implementations. The hardware implementation revealed to be highly resource expensive, mainly due the need of storing a whole video frame, but efficient in time, thanks to the parallelism feature of FPGA devices, executing quality score calculations between 20 and 40 times faster than a high-level language such as MATLAB. The fixed-point arithmetics revealed to be more efficient than the floating-point, specially regarding operation frequency.

La tesis se encarga de analizar, por un lado, la integración y el desarrollo de las nuevas tecnologías en la Administración de Justicia; y, por otro, los parámetros que constituyen la validez y eficacia del documento electrónico.
La primera cuestión se centra en la configuración de los Sistemas de Información de la Oficina Judicial y del Ministerio Fiscal, así como de la informatización de los Registros Civiles, donde el art. 230 LOPJ es la pieza clave. Se estudian sus programas, aplicaciones, la videoconferencia, los ficheros judiciales y las redes de telecomunicaciones que poseen la cobertura de la firma electrónica reconocida, donde cobran gran relevancia los convenios de colaboración tecnológica. La digitalización de las vistas quizá sea una de las cuestiones con más trascendencia, teniendo en cuenta que el juicio es el acto que culmina el proceso. Aunque no todos los proyectos adoptados en el ámbito de la e.justicia se han desarrollado de forma integral, ni han llegado a la totalidad de los órganos judiciales. El objetivo final es lograr una Justicia más ágil y de calidad, a lo cual aspira el Plan Estratégico de Modernización de la Justicia 2009-2012 aprobado recientemente.
En referencia a la segunda perspectiva, no cabe duda que el Ordenamiento jurídico y los tribunales, en el ámbito de la justicia material, otorgan plena validez y eficacia al documento electrónico. Nuestra línea de investigación se justifica porque cada vez son más los procesos que incorporan soportes electrónicos de todo tipo, ya sea al plantearse la acción o posteriormente como medio de prueba (art. 299.2 LEC). Entre otros temas examinamos el documento informático, la problemática que rodea al fax, los sistemas de videograbación y el contrato electrónico.
La tesi s'encarrega d'analitzar, per una part, la integració i el desenvolupament de les noves tecnologies dins l´Administració de Justícia; i, per l'altra, els paràmetres que constitueixen la validesa i l'eficàcia del document electrònic.
La primera qüestió es centra en la configuració dels Sistemes d´Informació de l´Oficina Judicial i del Ministeri Fiscal, així com de la informatització dels Registres Civils, on l'art. 230 LOPJ es la peça clau. S'estudien els seus programes, aplicacions, la videoconferència, el fitxers judicials i les xarxes de telecomunicacions que tenen la cobertura de la firma electrònica reconeguda, on cobren gran rellevància els convenis de col·laboració tecnològica. La digitalització de les vistes tal vegada sigui una de les qüestions amb més transcendència, tenint amb compte que el judici es l'acte que culmina el procés. Però no tots el projectes adoptats en l'àmbit de la e.justicia s'han desenvolupat d'una manera integral ni han arribat a la totalitat dels òrgans judicials. L'objectiu final es assolir una Justícia més àgil i de qualitat, al que aspira el Pla Estratègic de Modernització de la Justícia 2009-2012 aprovat recentment.
En referència a la segona perspectiva, no hi ha dubte que l´Ordenament jurídic i els tribunals, en l'àmbit de la justícia material, donen plena validesa i eficàcia al document electrònic. La nostra línia d'investigació es justifica perquè cada vegada son més el processos que incorporen suports electrònics de tot tipus, ja sigui quant es planteja l'acció o posteriorment como a medi de prova (art. 299.2 LEC). Entre altres temes examinem el document informàtic, la problemàtica que envolta al fax, els sistemes de videogravació i el contracte electrònic.
The thesis seeks to analyse, on the one hand, the integration and development of the new technologies in the Administration of Justice; and, on the other, the parameters which constitute the validity and efficiency of the electronic document.
The first question centres on the configuration of the Information Systems of the Judicial Office and the Public Prosecutor, as well as the computerisation of the Civil Registers, where the art. 230 LOPJ it's the part key. Their programmes, applications, the Video Conferencing, the judicial registers and the telecommunication networks which are covered by the recognised electronic signatures, are studied, where the agreements on technological collaboration gain great relevance. The digitalisation of evidence might perhaps be one of the questions with most consequence, bearing in mind that the judgment is the act by which the process is culminated. Although not all the projects adopted within the compass of e.justice have developed completely nor have reached all the judicial organs. The final objective is to achieve an agile, quality Justice, to which the recently approved Strategic Plan for the Modernisation of Justice aspires.
With reference to the second perspective, there is no doubt that the juridical Ordinance and the tribunals within the compass of material justice grant full validity and efficacy to the electronic document. Our line of investigation is justified because there are more and more processes which are sustained by electronic supports of all kinds, whether it be at the establishment of the action or later, as a proof of it (art. 299.2 LEC). Amongst other things, we examine the computerised document, the problems which surround the fax, the systems for video recording and the electronic contract.

Action understanding is key to automatically analyzing video content and thus is important for many real-world applications such as autonomous driving car, robot-assisted care, etc. Therefore, in the computer vision field, action understanding has been one of the fundamental research topics. Most conventional methods for action understanding are based on hand-crafted features. Like the recent advances seen in image classification, object detection, image captioning, etc, deep learning has become a popular approach for action understanding in video. However, there remain several important research challenges in developing deep learning based methods for understanding actions. This thesis focuses on the development of effective deep learning methods for solving three major challenges. Action detection at fine granularities in time: Previous work in deep learning based action understanding mainly focuses on exploring various backbone networks that are designed for the video-level action classification task. These did not explore the fine-grained temporal characteristics and thus failed to produce temporally precise estimation of action boundaries. In order to understand actions more comprehensively, it is important to detect actions at finer granularities in time. In Part I, we study both segment-level action detection and frame-level action detection. Segment-level action detection is usually formulated as the temporal action localization task, which requires not only recognizing action categories for the whole video but also localizing the start time and end time of each action instance. To this end, we propose an effective multi-stage framework called Segment-CNN consisting of three segment-based 3D ConvNets: (1) a proposal network identifies candidate segments that may contain actions; (2) a classification network learns one-vs-all action classification model to serve as initialization for the localization network; and (3) a localization network fine-tunes the learned classification network to localize each action instance. In another approach, frame-level action detection is effectively formulated as the per-frame action labeling task. We combine two reverse operations (i.e. convolution and deconvolution) into a joint Convolutional-De-Convolutional (CDC) filter, which simultaneously conducts downsampling in space and upsampling in time to jointly model both high-level semantics and temporal dynamics. We design a novel CDC network to predict actions at frame-level and the frame-level predictions can be further used to detect precise segment boundary for the temporal action localization task. Our method not only improves the state-of-the-art mean Average Precision (mAP) result on THUMOS’14 from 41.3% to 44.4% for the per-frame labeling task, but also improves mAP for the temporal action localization task from 19.0% to 23.3% on THUMOS’14 and from 16.4% to 23.8% on ActivityNet v1.3. Action detection in the constrained scenarios: The usual training process of deep learning models consists of supervision and data, which are not always available in reality. In Part II, we consider the scenarios of incomplete supervision and incomplete data. For incomplete supervision, we focus on the weakly-supervised temporal action localization task and propose AutoLoc which is the first framework that can directly predict the temporal boundary of each action instance with only the video-level annotations available during training. To enable the training of such a boundary prediction model, we design a novel Outer-Inner-Contrastive (OIC) loss to help discover the segment-level supervision and we prove that the OIC loss is differentiable to the underlying boundary prediction model. Our method significantly improves mAP on THUMOS14 from 13.7% to 21.2% and mAP on ActivityNet from 7.4% to 27.3%. For the scenario of incomplete data, we formulate a novel task called Online Detection of Action Start (ODAS) in streaming videos to enable detecting the action start time on the fly when a live video action is just starting. ODAS is important in many applications such as early alert generation to allow timely security or emergency response. Specifically, we propose three novel methods to address the challenges in training ODAS models: (1) hard negative samples generation based on Generative Adversarial Network (GAN) to distinguish ambiguous background, (2) explicitly modeling the temporal consistency between data around action start and data succeeding action start, and (3) adaptive sampling strategy to handle the scarcity of training data. Action understanding in the compressed domain: The mainstream action understanding methods including the aforementioned techniques developed by us require first decoding the compressed video into RGB image frames. This may result in significant cost in terms of storage and computation. Recently, researchers started to investigate how to directly perform action understanding in the compressed domain in order to achieve high efficiency while maintaining the state-of-the-art action detection accuracy. The key research challenge is developing effective backbone networks that can directly take data in the compressed domain as input. Our baseline is to take models developed for action understanding in the decoded domain and adapt them to attack the same tasks in the compressed domain. In Part III, we address two important issues in developing the backbone networks that exclusively operate in the compressed domain. First, compressed videos may be produced by different encoders or encoding parameters, but it is impractical to train a different compressed-domain action understanding model for each different format. We experimentally analyze the effect of video encoder variation and develop a simple yet effective training data preparation method to alleviate the sensitivity to encoder variation. Second, motion cues have been shown to be important for action understanding, but the motion vectors in compressed video are often very noisy and not discriminative enough for directly performing accurate action understanding. We develop a novel and highly efficient framework called DMC-Net that can learn to predict discriminative motion cues based on noisy motion vectors and residual errors in the compressed video streams. On three action recognition benchmarks, namely HMDB-51, UCF101 and a subset of Kinetics, we demonstrate that our DMC-Net can significantly shorten the performance gap between state-of-the-art compressed video based methods with and without optical flow, while being two orders of magnitude faster than the methods that use optical flow. By addressing the three major challenges mentioned above, we are able to develop more robust models for video action understanding and improve performance in various dimensions, such as (1) temporal precision, (2) required levels of supervision, (3) live video analysis ability, and finally (4) efficiency in processing compressed video. Our research has contributed significantly to advancing the state of the art of video action understanding and expanding the foundation for comprehensive semantic understanding of video content.

We demonstrate four novel multimodal methods for efficient video summarization and comprehensive cross-cultural news video understanding. First, For video quick browsing, we demonstrate a multimedia event recounting system. Based on nine people-oriented design principles, it summarizes YouTube-like videos into short visual segments (812sec) and textual words (less than 10 terms). In the 2013 Trecvid Multimedia Event Recounting competition, this system placed first in recognition time efficiency, while remaining above average in description accuracy. Secondly, we demonstrate the summarization of large amounts of online international news videos. In order to understand an international event such as Ebola virus, AirAsia Flight 8501 and Zika virus comprehensively, we present a novel and efficient constrained tensor factorization algorithm that first represents a video archive of multimedia news stories concerning a news event as a sparse tensor of order 4. The dimensions correspond to extracted visual memes, verbal tags, time periods, and cultures. The iterative algorithm approximately but accurately extracts coherent quad-clusters, each of which represents a significant summary of an important independent aspect of the news event. We give examples of quad-clusters extracted from tensors with at least 108 entries derived from international news coverage. We show the method is fast, can be tuned to give preferences to any subset of its four dimensions, and exceeds three existing methods in performance. Thirdly, noting that the co-occurrence of visual memes and tags in our summarization result is sparse, we show how to model cross-cultural visual meme influence based on normalized PageRank, which more accurately captures the rates at which visual memes are reposted in a specified time period in a specified culture. Lastly, we establish the correspondences of videos and text descriptions in different cultures by reliable visual cues, detect culture-specific tags for visual memes and then annotate videos in a cultural settings. Starting with any video with less text or no text in one culture (say, US), we select candidate annotations in the text of another culture (say, China) to annotate US video. Through analyzing the similarity of images annotated by those candidates, we can derive a set of proper tags from the viewpoints of another culture (China). We illustrate cultural-based annotation examples by segments of international news. We evaluate the generated tags by cross-cultural tag frequency, tag precision, and user studies.

Thesis (M.Sc.)--University of Alberta, 2009.
Title from PDF file main screen (viewed on Apr. 1, 2010). A thesis submitted to the Faculty of Graduate Studies and Research in partial fulfillment of the requirements for the degree of Master of Science, Department of Computing Science, University of Alberta. Includes bibliographical references.