Dissertations / Theses on the topic 'Hidden Markov support vector machine'

Aujourd’hui, les recherches portant sur le geste manquent de modèles génériques. Les spécialistes du geste doivent osciller entre une formalisation excessivement conceptuelle et une description purement visuelle du mouvement. Nous reprenons les concepts développés par le chorégraphe Rudolf Laban pour l’analyse de la danse classique contemporaine, et proposons leur extension afin d’élaborer un modèle générique du geste basé sur ses éléments expressifs. Nous présentons également deux corpus de gestes 3D que nous avons constitués. Le premier, ORCHESTRE-3D, se compose de gestes pré-segmentés de chefs d’orchestre enregistrés en répétition. Son annotation à l’aide d’émotions musicales est destinée à l’étude du contenu émotionnel de la direction musicale. Le deuxième corpus, HTI 2014-2015, propose des séquences d’actions variées de la vie quotidienne. Dans une première approche de reconnaissance dite « globale », nous définissons un descripteur qui se rapporte à l’entièreté du geste. Ce type de caractérisation nous permet de discriminer diverses actions, ainsi que de reconnaître les différentes émotions musicales que portent les gestes des chefs d’orchestre de notre base ORCHESTRE-3D. Dans une seconde approche dite « dynamique », nous définissons un descripteur de trame gestuelle (e.g. défini pour tout instant du geste). Les descripteurs de trame sont utilisés des poses-clés du mouvement, de sorte à en obtenir à tout instant une représentation simplifiée et utilisable pour reconnaître des actions à la volée. Nous testons notre approche sur plusieurs bases de geste, dont notre propre corpus HTI 2014-2015<br>Nowadays, researches dealing with gesture analysis suffer from a lack of unified mathematical models. On the one hand, gesture formalizations by human sciences remain purely theoretical and are not inclined to any quantification. On the other hand, the commonly used motion descriptors are generally purely intuitive, and limited to the visual aspects of the gesture. In the present work, we retain Laban Movement Analysis (LMA – originally designed for the study of dance movements) as a framework for building our own gesture descriptors, based on expressivity. Two datasets are introduced: the first one is called ORCHESTRE-3D, and is composed of pre-segmented orchestra conductors’ gestures, which have been annotated with the help of lexicon of musical emotions. The second one, HTI 2014-2015, comprises sequences of multiple daily actions. In a first experiment, we define a global feature vector based upon the expressive indices of our model and dedicated to the characterization of the whole gesture. This descriptor is used for action recognition purpose and to discriminate the different emotions of our orchestra conductors’ dataset. In a second approach, the different elements of our expressive model are used as a frame descriptor (e.g., describing the gesture at a given time). The feature space provided by such local characteristics is used to extract key poses of the motion. With the help of such poses, we obtain a per-frame sub-representation of body motions which is available for real-time action recognition purpose

The growth of multimedia content in recent years motivated the research on audio classification and content retrieval area. In this thesis, a general environmental audio classification and retrieval approach is proposed in which higher level semantic classes (outdoor, nature, meeting and violence) are obtained from lower level acoustic classes (emergency alarm, car horn, gun-shot, explosion, automobile, motorcycle, helicopter, wind, water, rain, applause, crowd and laughter). In order to classify an audio sample into acoustic classes, MPEG-7 audio features, Mel Frequency Cepstral Coefficients (MFCC) feature and Zero Crossing Rate (ZCR) feature are used with Hidden Markov Model (HMM) and Support Vector Machine (SVM) classifiers. Additionally, a new classification method is proposed using Genetic Algorithm (GA) for classification of semantic classes. Query by Example (QBE) and keyword-based query capabilities are implemented for content retrieval.

Communication through text messaging, SMS (Short Message Service), is nowadays a huge industry with billions of active users. Because of the huge userbase it has attracted many companies trying to market themselves through unsolicited messages in this medium in the same way as was previously done through email. This is such a common phenomenon that SMS spam has now become a plague in many countries. This report evaluates several established machine learning algorithms to see how well they can be applied to the problem of filtering unsolicited SMS messages. Each filter is mainly evaluated by analyzing the accuracy of the filters on stored message data. The report also discusses and compares requirements for hardware versus performance measured by how many messages that can be evaluated in a fixed amount of time. The results from the evaluation shows that a decision tree filter is the best choice of the filters evaluated. It has the highest accuracy as well as a high enough process rate of messages to be applicable. The decision tree filter which was found to be the most suitable for the task in this environment has been implemented. The accuracy in this new implementation is shown to be as high as the implementation used for the evaluation of this filter. Though the decision tree filter is shown to be the best choice of the filters evaluated it turned out the accuracy is not high enough to meet the specified requirements. It however shows promising results for further testing in this area by using improved methods on the best performing algorithms.

Computational biology is an interdisciplinary field that focuses on developing mathematical models and algorithms to interpret biological data so as to understand biological problems. With current high-throughput technology development, different types of biological data can be measured in a large scale, which calls for more sophisticated computational methods to analyze and interpret the data. In this dissertation research work, we propose novel methods to integrate, model and analyze multiple biological data, including microarray gene expression data, protein-DNA interaction data and protein-protein interaction data. These methods will help improve our understanding of biological systems. First, we propose a knowledge-guided multi-scale independent component analysis (ICA) method for biomarker identification on time course microarray data. Guided by a knowledge gene pool related to a specific disease under study, the method can determine disease relevant biological components from ICA modes and then identify biologically meaningful markers related to the specific disease. We have applied the proposed method to yeast cell cycle microarray data and Rsf-1-induced ovarian cancer microarray data. The results show that our knowledge-guided ICA approach can extract biologically meaningful regulatory modes and outperform several baseline methods for biomarker identification. Second, we propose a novel method for transcriptional regulatory network identification by integrating gene expression data and protein-DNA binding data. The approach is built upon a multi-level analysis strategy designed for suppressing false positive predictions. With this strategy, a regulatory module becomes increasingly significant as more relevant gene sets are formed at finer levels. At each level, a two-stage support vector regression (SVR) method is utilized to reduce false positive predictions by integrating binding motif information and gene expression data; a significance analysis procedure is followed to assess the significance of each regulatory module. The resulting performance on simulation data and yeast cell cycle data shows that the multi-level SVR approach outperforms other existing methods in the identification of both regulators and their target genes. We have further applied the proposed method to breast cancer cell line data to identify condition-specific regulatory modules associated with estrogen treatment. Experimental results show that our method can identify biologically meaningful regulatory modules related to estrogen signaling and action in breast cancer. Third, we propose a bootstrapping Markov Random Filed (MRF)-based method for subnetwork identification on microarray data by incorporating protein-protein interaction data. Methodologically, an MRF-based network score is first derived by considering the dependency among genes to increase the chance of selecting hub genes. A modified simulated annealing search algorithm is then utilized to find the optimal/suboptimal subnetworks with maximal network score. A bootstrapping scheme is finally implemented to generate confident subnetworks. Experimentally, we have compared the proposed method with other existing methods, and the resulting performance on simulation data shows that the bootstrapping MRF-based method outperforms other methods in identifying ground truth subnetwork and hub genes. We have then applied our method to breast cancer data to identify significant subnetworks associated with drug resistance. The identified subnetworks not only show good reproducibility across different data sets, but indicate several pathways and biological functions potentially associated with the development of breast cancer and drug resistance. In addition, we propose to develop network-constrained support vector machines (SVM) for cancer classification and prediction, by taking into account the network structure to construct classification hyperplanes. The simulation study demonstrates the effectiveness of our proposed method. The study on the real microarray data sets shows that our network-constrained SVM, together with the bootstrapping MRF-based subnetwork identification approach, can achieve better classification performance compared with conventional biomarker selection approaches and SVMs. We believe that the research presented in this dissertation not only provides novel and effective methods to model and analyze different types of biological data, the extensive experiments on several real microarray data sets and results also show the potential to improve the understanding of biological mechanisms related to cancers by generating novel hypotheses for further study.<br>Ph. D.

We investigate the unsupervised K-means clustering and the semi-supervised hidden Markov model (HMM) to automatically detect anomalous motion patterns in groups of people (crowds). Anomalous motion patterns are typically people merging into a dense group, followed by disturbances or threatening situations within the group. The application of K-means clustering and HMM are illustrated with datasets from four surveillance scenarios. The results indicate that by investigating the group of people in a systematic way with different K values, analyze cluster density, cluster quality and changes in cluster shape we can automatically detect anomalous motion patterns. The results correspond well with the events in the datasets. The results also indicate that very accurate detections of the people in the dense group would not be necessary. The clustering and HMM results will be very much the same also with some increased uncertainty in the detections.<br><p>Funding Agencies|Vinnova (Swedish Governmental Agency for Innovation Systems) under the VINNMER program||</p>

The detection of the Acoustic Events (AEs) naturally produced in a meeting room may help to describe the human and social activity. The automatic description of interactions between humans and environment can be useful for providing: implicit assistance to the people inside the room, context-aware and content-aware information requiring a minimum of human attention or interruptions, support for high-level analysis of the underlying acoustic scene, etc. On the other hand, the recent fast growth of available audio or audiovisual content strongly demands tools for analyzing, indexing, searching and retrieving the available documents. Given an audio document, the first processing step usually is audio segmentation (AS), i.e. the partitioning of the input audio stream into acoustically homogeneous regions which are labelled according to a predefined broad set of classes like speech, music, noise, etc. Acoustic event detection (AED) is the objective of this thesis work. A variety of features coming not only from audio but also from the video modality is proposed to deal with that detection problem in meeting-room and broadcast news domains. Two basic detection approaches are investigated in this work: a joint segmentation and classification using Hidden Markov Models (HMMs) with Gaussian Mixture Densities (GMMs), and a detection-by-classification approach using discriminative Support Vector Machines (SVMs). For the first case, a fast one-pass-training feature selection algorithm is developed in this thesis to select, for each AE class, the subset of multimodal features that shows the best detection rate. AED in meeting-room environments aims at processing the signals collected by distant microphones and video cameras in order to obtain the temporal sequence of (possibly overlapped) AEs that have been produced in the room. When applied to interactive seminars with a certain degree of spontaneity, the detection of acoustic events from only the audio modality alone shows a large amount of errors, which is mostly due to the temporal overlaps of sounds. This thesis includes several novelties regarding the task of multimodal AED. Firstly, the use of video features. Since in the video modality the acoustic sources do not overlap (except for occlusions), the proposed features improve AED in such rather spontaneous scenario recordings. Secondly, the inclusion of acoustic localization features, which, in combination with the usual spectro-temporal audio features, yield a further improvement in recognition rate. Thirdly, the comparison of feature-level and decision-level fusion strategies for the combination of audio and video modalities. In the later case, the system output scores are combined using two statistical approaches: weighted arithmetical mean and fuzzy integral. On the other hand, due to the scarcity of annotated multimodal data, and, in particular, of data with temporal sound overlaps, a new multimodal database with a rich variety of meeting-room AEs has been recorded and manually annotated, and it has been made publicly available for research purposes.<br>La detecció d'esdeveniments acústics (Acoustic Events -AEs-) que es produeixen naturalment en una sala de reunions pot ajudar a descriure l'activitat humana i social. La descripció automàtica de les interaccions entre els éssers humans i l'entorn pot ser útil per a proporcionar: ajuda implícita a la gent dins de la sala, informació sensible al context i al contingut sense requerir gaire atenció humana ni interrupcions, suport per a l'anàlisi d'alt nivell de l'escena acústica, etc. La detecció i la descripció d'activitat és una funcionalitat clau de les interfícies perceptives que treballen en entorns de comunicació humana com sales de reunions. D'altra banda, el recent creixement ràpid del contingut audiovisual disponible requereix l'existència d'eines per a l'anàlisi, indexació, cerca i recuperació dels documents existents. Donat un document d'àudio, el primer pas de processament acostuma a ser la seva segmentació (Audio Segmentation (AS)), és a dir, la partició de la seqüència d'entrada d'àudio en regions acústiques homogènies que s'etiqueten d'acord amb un conjunt predefinit de classes com parla, música, soroll, etc. De fet, l'AS pot ser vist com un cas particular de la detecció d’esdeveniments acústics, i així es fa en aquesta tesi. La detecció d’esdeveniments acústics (Acoustic Event Detection (AED)) és un dels objectius d'aquesta tesi. Es proposa tot una varietat de característiques que provenen no només de l'àudio, sinó també de la modalitat de vídeo, per fer front al problema de la detecció en dominis de sala de reunions i de difusió de notícies. En aquest treball s'investiguen dos enfocaments bàsics de detecció: 1) la realització conjunta de segmentació i classificació utilitzant models de Markov ocults (Hidden Markov Models (HMMs)) amb models de barreges de gaussianes (Gaussian Mixture Models (GMMs)), i 2) la detecció per classificació utilitzant màquines de vectors suport (Support Vector Machines (SVM)) discriminatives. Per al primer cas, en aquesta tesi es desenvolupa un algorisme de selecció de característiques ràpid d'un sol pas per tal de seleccionar, per a cada AE, el subconjunt de característiques multimodals que aconsegueix la millor taxa de detecció. L'AED en entorns de sales de reunió té com a objectiu processar els senyals recollits per micròfons distants i càmeres de vídeo per tal d'obtenir la seqüència temporal dels (possiblement superposats) esdeveniments acústics que s'han produït a la sala. Quan s'aplica als seminaris interactius amb un cert grau d'espontaneïtat, la detecció d'esdeveniments acústics a partir de només la modalitat d'àudio mostra una gran quantitat d'errors, que és sobretot a causa de la superposició temporal dels sons. Aquesta tesi inclou diverses contribucions pel que fa a la tasca d'AED multimodal. En primer lloc, l'ús de característiques de vídeo. Ja que en la modalitat de vídeo les fonts acústiques no se superposen (exceptuant les oclusions), les característiques proposades Resum iv milloren la detecció en els enregistraments en escenaris de caire espontani. En segon lloc, la inclusió de característiques de localització acústica, que, en combinació amb les característiques habituals d'àudio espectrotemporals, signifiquen nova millora en la taxa de reconeixement. En tercer lloc, la comparació d'estratègies de fusió a nivell de característiques i a nivell de decisions, per a la utilització combinada de les modalitats d'àudio i vídeo. En el darrer cas, les puntuacions de sortida del sistema es combinen fent ús de dos mètodes estadístics: la mitjana aritmètica ponderada i la integral difusa. D'altra banda, a causa de l'escassetat de dades multimodals anotades, i, en particular, de dades amb superposició temporal de sons, s'ha gravat i anotat manualment una nova base de dades multimodal amb una rica varietat d'AEs de sala de reunions, i s'ha posat a disposició pública per a finalitats d'investigació. Per a la segmentació d'àudio en el domini de difusió de notícies, es proposa una arquitectura jeràrquica de sistema, que agrupa apropiadament un conjunt de detectors, cada un dels quals correspon a una de les classes acústiques d'interès. S'han desenvolupat dos sistemes diferents de SA per a dues bases de dades de difusió de notícies: la primera correspon a gravacions d'àudio del programa de debat Àgora del canal de televisió català TV3, i el segon inclou diversos segments d'àudio del canal de televisió català 3/24 de difusió de notícies. La sortida del primer sistema es va utilitzar com a primera etapa dels sistemes de traducció automàtica i de subtitulat del projecte Tecnoparla, un projecte finançat pel govern de la Generalitat en el que es desenvoluparen diverses tecnologies de la parla per extreure tota la informació possible del senyal d'àudio. El segon sistema d'AS, que és un sistema de detecció jeràrquica basat en HMM-GMM amb selecció de característiques, ha obtingut resultats competitius en l'avaluació de segmentació d'àudio Albayzín2010. Per acabar, val la pena esmentar alguns resultats col·laterals d’aquesta tesi. L’autor ha sigut responsable de l'organització de l'avaluació de sistemes de segmentació d'àudio dins de la campanya Albayzín-2010 abans esmentada. S'han especificat les classes d’esdeveniments, les bases de dades, la mètrica i els protocols d'avaluació utilitzats, i s'ha realitzat una anàlisi posterior dels sistemes i els resultats presentats pels vuit grups de recerca participants, provinents d'universitats espanyoles i portugueses. A més a més, s'ha implementat en la sala multimodal de la UPC un sistema de detecció d'esdeveniments acústics per a dues fonts simultànies, basat en HMM-GMM, i funcionant en temps real, per finalitats de test i demostració.

Genetic regulatory networks (GRN) offer a useful model for clinical biology. Specifically, such networks capture interactions among genes, proteins, and other metabolic factors. Unfortunately, it is difficult to understand and predict the behavior of networks that are of realistic size and complexity. In this dissertation, behavior refers to the trajectory of a state, through a series of state transitions over time, to an attractor in the network. This project assumes asynchronous Boolean networks, implying that a state may transition to more than one attractor. The goal of this project is to efficiently identify a network's set of attractors and to predict the likelihood with which an arbitrary state leads to each of the network’s attractors. These probabilities will be represented using a fuzzy membership vector. Predicting fuzzy membership vectors using machine learning techniques may address the intractability posed by networks of realistic size and complexity. Modeling and simulation can be used to provide the necessary training sets for machine learning methods to predict fuzzy membership vectors. The experiments comprise several GRNs, each represented by a set of output classes. These classes consist of thresholds τ and ¬τ, where τ = [τlaw,τhigh]; state s belongs to class τ if the probability of its transitioning to attractor 􀜣 belongs to the range [τlaw,τhigh]; otherwise it belongs to class ¬τ. Finally, each machine learning classifier was trained with the training sets that was previously collected. The objective is to explore methods to discover patterns for meaningful classification of states in realistically complex regulatory networks. The research design took a GRN and a machine learning method as input and produced output class < Ατ > and its negation ¬ < Ατ >. For each GRN, attractors were identified, data was collected by sampling each state to create fuzzy membership vectors, and machine learning methods were trained to predict whether a state is in a healthy attractor or not. For T-LGL, SVMs had the highest accuracy in predictions (between 93.6% and 96.9%) and precision (between 94.59% and 97.87%). However, naive Bayesian classifiers had the highest recall (between 94.71% and 97.78%). This study showed that all experiments have extreme significance with pvalue < 0.0001. The contribution this research offers helps clinical biologist to submit genetic states to get an initial result on their outcomes. For future work, this implementation could use other machine learning classifiers such as xgboost or deep learning methods. Other suggestions offered are developing methods that improves the performance of state transition that allow for larger training sets to be sampled.

A good image representation is vital for image comparision and classification; it may affect the classification accuracy and efficiency. The purpose of this thesis was to explore novel and appropriate image representations. Another aim was to investigate these representations for image classification. Finally, novel features were examined for improving image classification accuracy. Images of interest to this thesis were textile design images. The motivation of analysing textile design images is to help designers browse images, fuel their creativity, and improve their design efficiency. In recent years, bag-of-words model has been shown to be a good base for image representation, and there have been many attempts to go beyond this representation. Bag-of-words models have been used frequently in the classification of image data, due to good performance and simplicity. “Words” in images can have different definitions and are obtained through steps of feature detection, feature description, and codeword calculation. The model represents an image as an orderless collection of local features. However, discarding the spatial relationships of local features limits the power of this model. This thesis exploited novel image representations, bag of shapes and region label graphs models, which were based on bag-of-words model. In both models, an image was represented by a collection of segmented regions, and each region was described by shape descriptors. In the latter model, graphs were constructed to capture the spatial information between groups of segmented regions and graph features were calculated based on some graph theory. Novel elements include use of MRFs to extract printed designs and woven patterns from textile images, utilisation of the extractions to form bag of shapes models, and construction of region label graphs to capture the spatial information. The extraction of textile designs was formulated as a pixel labelling problem. Algorithms for MRF optimisation and re-estimation were described and evaluated. A method for quantitative evaluation was presented and used to compare the performance of MRFs optimised using alpha-expansion and iterated conditional modes (ICM), both with and without parameter re-estimation. The results were used in the formation of the bag of shapes and region label graphs models. Bag of shapes model was a collection of MRFs' segmented regions, and the shape of each region was described with generic Fourier descriptors. Each image was represented as a bag of shapes. A simple yet competitive classification scheme based on nearest neighbour class-based matching was used. Classification performance was compared to that obtained when using bags of SIFT features. To capture the spatial information, region label graphs were constructed to obtain graph features. Regions with the same label were treated as a group and each group was associated uniquely with a vertex in an undirected, weighted graph. Each region group was represented as a bag of shape descriptors. Edges in the graph denoted either the extent to which the groups' regions were spatially adjacent or the dissimilarity of their respective bags of shapes. Series of unweighted graphs were obtained by removing edges in order of weight. Finally, an image was represented using its shape descriptors along with features derived from the chromatic numbers or domination numbers of the unweighted graphs and their complements. Linear SVM classifiers were used for classification. Experiments were implemented on data from Liberty Art Fabrics, which consisted of more than 10,000 complicated images mainly of printed textile designs and woven patterns. Experimental data was classified into seven classes manually by assigning each image a text descriptor based on content or design type. The seven classes were floral, paisley, stripe, leaf, geometric, spot, and check. The result showed that reasonable and interesting regions were obtained from MRF segmentation in which alpha-expansion with parameter re-estimation performs better than alpha-expansion without parameter re-estimation or ICM. This result was not only promising for textile CAD (Computer-Aided Design) to redesign the textile image, but also for image representation. It was also found that bag of shapes model based on MRF segmentation can obtain comparable classification accuracy with bag of SIFT features in the framework of nearest neighbour class-based matching. Finally, the result indicated that incorporation of graph features extracted by constructing region label graphs can improve the classification accuracy compared to both bag of shapes model and bag of SIFT models.

Dans le cadre de cette thèse nous nous sommes intéressés à l'application des outils d'apprentissage statistique aux problèmes d'extraction et de suivi d'objets dans le contexte de la surveillance vidéo par des caméras statiques. Notre étude se déroule en trois phases : la première consiste à détecter l'ensemble des objets en mouvement par l'analyse des variations de luminosité au cours du temps. Nous avons proposé pour cela une méthode de détection de tels objets dans la continuité des travaux de Grimson et Stauffer à la base de la modélisation de la luminance des pixels par des mélanges des lois Gaussiennes. Nous proposons une règle de décision statistique au sens du MAP simplifiée pour satisfaire les contraintes temps-réel. Une régularisation des résultats de la détection par modélisation des champs des étiquettes en tant que champs de Markov nous a permis d'obtenir des masques de mouvement les plus complets aussi bien au niveau spatial que temporel. Suite à cette première analyse, un deuxième processus de détection est exécuté en vue de l'identification d'un ou plusieurs objets d'intérêt. L'ensemble de ces objets est prédéfini au préalable par un opérateur. Dans le cadre de ce travail, nous avons envisagé une application concrète, celle de la détection de visages, tout en considérant des solutions statistiques de classification suffisamment généralistes ainsi qu'une étude des espaces de représentation adéquats. Nous avons étudié et déployé un classifieur à base de Support Vector Machine -SVM-. Afin d'optimiser les temps de calcul, nous avons proposé une méthode d'élimination des vecteurs de support linéairement dépendants ainsi qu'un schéma de détection de l'objet d'intérêt uniquement dans les zones de mouvement précédemment détectées. Enfin, la troisième contribution de cette thèse consiste en un outil de suivi d'objet appris avec un modèle de mouvement affine complet de premier ordre et ceci à la base du même formalisme SVM. Ici les paramètres de classifieur appris à l'étape de la détection sont réutilisés pour le suivi. Ce modèle permet ainsi de gérer les mouvements complexes des objets filmés dans des environnements naturels.

碩士<br>國立中央大學<br>資訊管理研究所<br>94<br>Various malicious programs, such as Trojan horse and backdoor, have become popular on the Internet in recent years. More and more automated penetration testing tools appear and now less background knowledge of attack is needed than before. As a result, the responsibility of computer is transferred to the host-based intrusion detection systems. Our research mainly combines Hidden Markov Model and Support Vector Machine and proposes a host-based anomaly detection system under Windows platforms. We use Windows Native Application Interface (API) sequences to establish the program normal behavior model. This kind of data has a significant characteristic that is the order of API appearing sequence. So we utilize the Hidden Markov Model that is good at expressing dynamic sequences relation to describe the probability relation of order between Windows Native APIs. After obtaining the hidden state sequences of Native API sequences by Hidden Markov Model, we put it into Support Vector Machine to train normal behavior of programs. If our prototype system detects the state of program belonging to the anomaly, we can inform users about the anomalous behavior of the program. A prototype system is developed by us using the proposed method. We did several experiments to evaluate the performance of this system. The experiments use the dataset of the New Mexico University and the data of the Windows Native API dataset collected by ourselves. The results of experiments prove the effectiveness of the combination of the Hidden Markov Model and Support Vector Machine that can distinguish anomalous program behavior from normal program behavior.

Our goal is to develop a state-of-the-art predictor with an intuitive and biophysically-motivated energy model through the use of Hidden Markov Support Vector Machines (HM-SVMs), a recent innovation in the field of machine learning. We focus on the prediction of alpha helices in proteins and show that using HM-SVMs, a simple 7-state HMM with 302 parameters can achieve a Q_alpha value of 77.6% and a SOV_alpha value of 73.4%. We briefly describe how our method can be generalized to predicting beta strands and sheets.

Hidden Markov Models (HMMs) are, undoubtedly, the most employed core technique for Automatic Speech Recognition (ASR). Nevertheless, we are still far from achieving high‐performance ASR systems. Some alternative approaches, most of them based on Artificial Neural Networks (ANNs), were proposed during the late eighties and early nineties. Some of them tackled the ASR problem using predictive ANNs, while others proposed hybrid HMM/ANN systems. However, despite some achievements, nowadays, the dependency on Hidden Markov Models is a fact. During the last decade, however, a new tool appeared in the field of machine learning that has proved to be able to cope with hard classification problems in several fields of application: the Support Vector Machines (SVMs). The SVMs are effective discriminative classifiers with several outstanding characteristics, namely: their solution is that with maximum margin; they are capable to deal with samples of a very higher dimensionality; and their convergence to the minimum of the associated cost function is guaranteed. In this work a novel approach based upon probabilistic kernels in support vector machines have been attempted for speech data classification. The classification accuracy in case of support vector classification depends upon the kernel function used which in turn depends upon the data set in hand. But still as of now there is no way to know a priori which kernel will give us best results The kernel used in this work tries to normalize the time dimension by fitting a probability distribution over individual data points which normalizes the time dimension inherent to speech signals which facilitates the use of support vector machines since it acts on static data only. The divergence between these probability distributions fitted over individual speech utterances is used to form the kernel matrix. Vowel Classification, Isolated Word Recognition (Digit Recognition), have been attempted and results are compared with state of art systems.

Magister Scientiae - MSc<br>Research has shown that five fundamental parameters are required to recognize any sign language gesture: hand shape, hand motion, hand location, hand orientation and facial expressions. The South African Sign Language (SASL) research group at the University of the Western Cape (UWC) has created several systems to recognize sign language gestures using single parameters. These systems are, however, limited to a vocabulary size of 20 – 23 signs, beyond which the recognition accuracy is expected to decrease. The first aim of this research is to investigate the use of two parameters – hand motion and hand shape – to recognise a larger vocabulary of SASL gestures at a high accuracy. Also, the majority of related work in the field of sign language gesture recognition using these two parameters makes use of Hidden Markov Models (HMMs) to classify gestures. Hidden Markov Support Vector Machines (HM-SVMs) are a relatively new technique that make use of Support Vector Machines (SVMs) to simulate the functions of HMMs. Research indicates that HM-SVMs may perform better than HMMs in some applications. To our knowledge, they have not been applied to the field of sign language gesture recognition. This research compares the use of these two techniques in the context of SASL gesture recognition. The results indicate that, using two parameters results in a 15% increase in accuracy over the use of a single parameter. Also, it is shown that HM-SVMs are a more accurate technique than HMMs, generally performing better or at least as good as HMMs.

The automotive industry is growing bigger each year. The central concern for any automotive company is driver and passenger safety. Many automotive companies have developed driver assistance systems, to help the driver and to ensure driver safety. These systems include adaptive cruise control, lane departure warning, lane change assistance, collision avoidance, night vision, automatic parking, traffic sign recognition, and driver fatigue detection. In this thesis, we aim to build a driver fatigue detection system that advances the research in this area. Using vision in detecting driver fatigue is commonly the key part for driver fatigue detection systems. We have decided to investigate different direction. We examine the driver's voice, heart rate, and driving performance to assess fatigue level. The system consists of three main modules: the audio module, the heart rate and other signals module, and the Bayesian network module. The audio module analyzes an audio recording of a driver and tries to estimate the level of fatigue for the driver. A Voice Activity Detection (VAD) module is used to extract driver speech from the audio recording. Mel-Frequency Cepstrum Coefficients, (MFCC) features are extracted from the speech signal, and then Support Vector Machines (SVM) and Hidden Markov Models (HMM) classifiers are used to detect driver fatigue. Both classifiers are tuned for best performance, and the performance of both classifiers is reported and compared. The heart rate and other signals module uses heart rate, steering wheel position, and the positions of the accelerator, brake, and clutch pedals to detect the level of fatigue. These signals' sample rates are then adjusted to match, allowing simple features to be extracted from the signals, and SVM and HMM classifiers are used to detect fatigue level. The performance of both classifiers is reported and compared. Bayesian networks' abilities to capture dependencies and uncertainty make them a sound choice to perform the data fusion. Prior information (Day/Night driving and previous decision) is also incorporated into the network to improve the final decision. The accuracies of the audio and heart rate and other signals modules are used to calculate certain CPTs for the Bayesian network, while the rest of the CPTs are calculated subjectively. The inference queries are calculated using the variable elimination algorithm. For those time steps where the audio module decision is absent, a window is defined and the last decision within this window is used as a current decision. The performance of the system is assessed based on the average accuracy per second. A dataset was built to train and test the system. The experimental results show that the system is very promising. The performance of the system was assessed based on the average accuracy per second; the total accuracy of the system is 90.5%. The system design can be easily improved by easily integrating more modules into the Bayesian network.