Made available in DSpace on 2014-12-17T14:54:49Z (GMT). No. of bitstreams: 1 HelianaBS_da_capa_ate_cap4.pdf: 2361373 bytes, checksum: 3e1e43e8ba1aadc274663b8b8e3de72f (MD5) Previous issue date: 2008-02-22 / Conselho Nacional de Desenvolvimento Cient?fico e Tecnol?gico / The skin cancer is the most common of all cancers and the increase of its incidence must, in part, caused by the behavior of the people in relation to the exposition to the sun. In Brazil, the non-melanoma skin cancer is the most incident in the majority of the regions. The dermatoscopy and videodermatoscopy are the main types of examinations for the diagnosis of dermatological illnesses of the skin. The field that involves the use of computational tools to help or follow medical diagnosis in dermatological injuries is seen as very recent. Some methods had been proposed for automatic classification of pathology of the skin using images. The present work has the objective to present a new intelligent methodology for analysis and classification of skin cancer images, based on the techniques of digital processing of images for extraction of color characteristics, forms and texture, using Wavelet Packet Transform (WPT) and learning techniques called Support Vector Machine (SVM). The Wavelet Packet Transform is applied for extraction of texture characteristics in the images. The WPT consists of a set of base functions that represents the image in different bands of frequency, each one with distinct resolutions corresponding to each scale. Moreover, the characteristics of color of the injury are also computed that are dependants of a visual context, influenced for the existing colors in its surround, and the attributes of form through the Fourier describers. The Support Vector Machine is used for the classification task, which is based on the minimization principles of the structural risk, coming from the statistical learning theory. The SVM has the objective to construct optimum hyperplanes that represent the separation between classes. The generated hyperplane is determined by a subset of the classes, called support vectors. For the used database in this work, the results had revealed a good performance getting a global rightness of 92,73% for melanoma, and 86% for non-melanoma and benign injuries. The extracted describers and the SVM classifier became a method capable to recognize and to classify the analyzed skin injuries / O c?ncer de pele ? o mais comum de todos os c?nceres e o aumento da sua incid?ncia deve-se, em parte, ao comportamento das pessoas em rela??o ? exposi??o ao sol. No Brasil, o c?ncer de pele n?o melanoma ? o mais incidente na maioria das regi?es. A dermatoscopia e ideodermatoscopia s?o os principais tipos de exames para o diagn?stico de doen?as da pele dermatol?gicas. O campo que envolve o uso de ferramentas computacionais para o aux?lio ou acompanhamento do diagn?stico m?dico em les?es dermatol?gicas ainda ? visto como muito recente. V?rios m?todos foram propostos para classifica??o autom?tica de patologias da pele utilizando imagens. O presente trabalho tem como objetivo apresentar uma nova metodologia inteligente para an?lise e classifica??o de imagens de c?ncer de pele, baseada nas t?cnicas de processamento digital de imagens para extra??o de caracter?sticas de cor, forma e textura, utilizando a Transformada Wavelet Packet (TWP) e a t?cnicas de aprendizado de m?quina denominada M?quina de Vetor de Suporte (SVM Support Vector Machine). A Transformada Wavelet Packet ? aplicada para extra??o de caracter?sticas de textura nas imagens. Esta consiste de um conjunto de fun??es base que representa a imagem em diferentes bandas de freq??ncia, cada uma com resolu??es distintas correspondente a cada escala. Al?m disso, s?o calculadas tamb?m as caracter?sticas de cor da les?o que s?o dependentes de um contexto visual, influenciada pelas cores existentes em sua volta, e os atributos de forma atrav?s dos descritores de Fourier. Para a tarefa de classifica??o ? utilizado a M?quina de Vetor de Suporte, que baseia-se nos princ?pios da minimiza??o do risco estrutural, proveniente da teoria do aprendizado estat?stico. A SVM tem como objetivo construir hiperplanos ?timos que apresentem a maior margem de separa??o entre classes. O hiperplano gerado ? determinado por um subconjunto dos pontos das classes, chamado vetores de suporte. Para o banco de dados utilizado neste trabalho, os resultados apresentaram um bom desempenho obtendo um acerto global de 92,73% para melanoma, e 86% para les?es n?o-melanoma e benigna. O potencial dos descritores extra?dos aliados ao classificador SVM tornou o m?todo capaz de reconhecer e classificar as les?es analisadas

Traditionally, local shape descriptors or color and texture based descriptors are used to describe the content of images. Although, these solutions achieving good results, they are not able to distinguish scenes that contain objects with the same colors, but with a different spatial organization or do not supports partial matching. In this work we focus on a particular case of the partial matching that is to find individual objects in images that contain various objects. Since the color is one of the most visually distinguishable properties, we propose a new descriptor based only on color able to find pictures of objects that are contained in other images. Although our descriptor has shown better results when compared to related works, this new color descriptor is not able to discriminate objects topologically different but having the same colors. To overcome this problem, we also propose a new approach to the partial matching of images that combine color and topological features on a single descriptor. This new descriptor, first performs a simplification process of the original image, which identifies the color regions that make up the image. Then, we represent the spatial information among the color regions using a topological graph, where vertices represent the color regions and the edges represent the spatial connections between them. To calculate the descriptor from this graph representation, we use the spectral theory of graphs, avoiding the need to make a direct comparison between graphs. To support the partial matching, we propose a decomposition of the main graph into several subgraphs, and also calculate descriptors for these subgraphs. / Tradicionalmente, descritores de forma, ou descritores baseados em cor e textura, são utilizados para descrever o conteúdo visual das imagens. Embora essas abordagens apresentem bons resultados, elas não são capazes de diferenciar adequadamente imagens que contêm objetos com as mesmas cores, mas com organização espacial diferente ou não suportam a pesquisa parcial de imagens. Neste trabalho focamos em um caso particular da pesquisa parcial de imagens, que é encontrar objetos em imagens que contenham vários objetos, não deixando de lado a pesquisa total (encontrar imagens similares à original). Dado que a cor é uma das propriedades visuais mais discriminativas, propomos um novo descritor baseado somente em cor capaz de encontrar imagens de objetos que estão contidos em outras imagens. Embora tenha apresentado melhores resultados quando comparado a trabalhos correlatos, esse novo descritor de cor não é capaz de discriminar objetos topologicamente diferentes mas que possuam as mesmas cores. Com o intuito de resolver esse problema, também propomos uma nova abordagem para a recuperação parcial de imagens que combina características topológicas e de cor em um único descritor. Esse novo descritor primeiramente realiza um processo de simplificação da imagem original, onde são identificadas as regiões de cor que compõem a imagem. Após esse processo de simplificação, a organização espacial das regiões de cor previamente identificadas é representada por meio de um grafo topológico, onde os vértices representam as regiões de cor e as arestas representam as conexões entre essas regiões. O descritor topológico é então calculado a partir do grafo de topologia utilizando a teoria espectral de grafos, evitando a necessidade de se realizar uma comparação direta entre grafos. Para suportar a pesquisa parcial de imagens, é realizada uma decomposição do grafo principal em diversos subgrafos. / Mestre em Ciência da Computação

Cette thèse est consacrée au problème de la reconnaissance visuelle des objets basé sur l'ordinateur, qui est devenue un sujet de recherche très populaire et important ces dernières années grâce à ses nombreuses applications comme l'indexation et la recherche d'image et de vidéo , le contrôle d'accès de sécurité, la surveillance vidéo, etc. Malgré beaucoup d'efforts et de progrès qui ont été fait pendant les dernières années, il reste un problème ouvert et est encore considéré comme l'un des problèmes les plus difficiles dans la communauté de vision par ordinateur, principalement en raison des similarités entre les classes et des variations intra-classe comme occlusion, clutter de fond, les changements de point de vue, pose, l'échelle et l'éclairage. Les approches populaires d'aujourd'hui pour la reconnaissance des objets sont basé sur les descripteurs et les classiffieurs, ce qui généralement extrait des descripteurs visuelles dans les images et les vidéos d'abord, et puis effectue la classification en utilisant des algorithmes d'apprentissage automatique sur la base des caractéristiques extraites. Ainsi, il est important de concevoir une bonne description visuelle, qui devrait être à la fois discriminatoire et efficace à calcul, tout en possédant certaines propriétés de robustesse contre les variations mentionnées précédemment. Dans ce contexte, l’objectif de cette thèse est de proposer des contributions novatrices pour la tâche de la reconnaissance visuelle des objets, en particulier de présenter plusieurs nouveaux descripteurs visuelles qui représentent effectivement et efficacement le contenu visuel d’image et de vidéo pour la reconnaissance des objets. Les descripteurs proposés ont l'intention de capturer l'information visuelle sous aspects différents. Tout d'abord, nous proposons six caractéristiques LBP couleurs de multi-échelle pour traiter les défauts principaux du LBP original, c'est-à-dire, le déffcit d'information de couleur et la sensibilité aux variations des conditions d'éclairage non-monotoniques. En étendant le LBP original à la forme de multi-échelle dans les différents espaces de couleur, les caractéristiques proposées non seulement ont plus de puissance discriminante par l'obtention de plus d'information locale, mais possèdent également certaines propriétés d'invariance aux différentes variations des conditions d’éclairage. En plus, leurs performances sont encore améliorées en appliquant une stratégie de l'image division grossière à fine pour calculer les caractéristiques proposées dans les blocs d'image afin de coder l'information spatiale des structures de texture. Les caractéristiques proposées capturent la distribution mondiale de l’information de texture dans les images. Deuxièmement, nous proposons une nouvelle méthode pour réduire la dimensionnalité du LBP appelée la combinaison orthogonale de LBP (OC-LBP). Elle est adoptée pour construire un nouveau descripteur local basé sur la distribution en suivant une manière similaire à SIFT. Notre objectif est de construire un descripteur local plus efficace en remplaçant l'information de gradient coûteux par des patterns de texture locales dans le régime du SIFT. Comme l'extension de notre première contribution, nous étendons également le descripteur OC-LBP aux différents espaces de couleur et proposons six descripteurs OC-LBP couleurs pour améliorer la puissance discriminante et la propriété d'invariance photométrique du descripteur basé sur l'intensité. Les descripteurs proposés capturent la distribution locale de l’information de texture dans les images. Troisièmement, nous introduisons DAISY, un nouveau descripteur local rapide basé sur la distribution de gradient, dans le domaine de la reconnaissance visuelle des objets. [...] / This thesis is dedicated to the problem of machine-based visual object recognition, which has become a very popular and important research topic in recent years because of its wide range of applications such as image/video indexing and retrieval, security access control, video monitoring, etc. Despite a lot of e orts and progress that have been made during the past years, it remains an open problem and is still considered as one of the most challenging problems in computer vision community, mainly due to inter-class similarities and intra-class variations like occlusion, background clutter, changes in viewpoint, pose, scale and illumination. The popular approaches for object recognition nowadays are feature & classifier based, which typically extract visual features from images/videos at first, and then perform the classification using certain machine learning algorithms based on the extracted features. Thus it is important to design good visual description, which should be both discriminative and computationally efficient, while possessing some properties of robustness against the previously mentioned variations. In this context, the objective of this thesis is to propose some innovative contributions for the task of visual object recognition, in particular to present several new visual features / descriptors which effectively and efficiently represent the visual content of images/videos for object recognition. The proposed features / descriptors intend to capture the visual information from different aspects. Firstly, we propose six multi-scale color local binary pattern (LBP) features to deal with the main shortcomings of the original LBP, namely deficiency of color information and sensitivity to non-monotonic lighting condition changes. By extending the original LBP to multi-scale form in different color spaces, the proposed features not only have more discriminative power by obtaining more local information, but also possess certain invariance properties to different lighting condition changes. In addition, their performances are further improved by applying a coarse-to-fine image division strategy for calculating the proposed features within image blocks in order to encode spatial information of texture structures. The proposed features capture global distribution of texture information in images. Secondly, we propose a new dimensionality reduction method for LBP called the orthogonal combination of local binary patterns (OC-LBP), and adopt it to construct a new distribution-based local descriptor by following a way similar to SIFT.Our goal is to build a more efficient local descriptor by replacing the costly gradient information with local texture patterns in the SIFT scheme. As the extension of our first contribution, we also extend the OC-LBP descriptor to different color spaces and propose six color OC-LBP descriptors to enhance the discriminative power and the photometric invariance property of the intensity-based descriptor. The proposed descriptors capture local distribution of texture information in images. Thirdly, we introduce DAISY, a new fast local descriptor based on gradient distribution, to the domain of visual object recognition.

Submitted by Fabio Sobreira Campos da Costa (fabio.sobreira@ufpe.br) on 2017-02-23T14:02:54Z No. of bitstreams: 2 license_rdf: 1232 bytes, checksum: 66e71c371cc565284e70f40736c94386 (MD5) tese-completa-Sidney_Lima_v21.pdf: 4757211 bytes, checksum: 205170db8b002cc2ab72255ab77628a3 (MD5) / Made available in DSpace on 2017-02-23T14:02:54Z (GMT). No. of bitstreams: 2 license_rdf: 1232 bytes, checksum: 66e71c371cc565284e70f40736c94386 (MD5) tese-completa-Sidney_Lima_v21.pdf: 4757211 bytes, checksum: 205170db8b002cc2ab72255ab77628a3 (MD5) Previous issue date: 2016-02-25 / REUNI / O câncer de mama é a principal causa de morte de mulheres adultas por câncer no mundo. Do ponto de vista clínico, a mamografia é ainda a mais efetiva tecnologia de diagnóstico do câncer de mama, dada a grande difusão de uso e interpretações dessas imagens. De acordo com o estado da arte da classificação de lesões em mamogramas, as wavelets têm apresentado os melhores resultados do ponto de vista da taxa de classificação, quando utilizadas como etapa de préprocessamento que decompõe a imagem original em imagens de detalhes (verticais, horizontais e diagonais) e aproximações para, a partir dessas imagens componentes, serem extraídos atributos de textura. Neste trabalho, propõe-se a Decomposição baseada em Aproximações Morfológicas em regiões de interesse em mamogramas. O método proposto tem por base a decomposição inspirada em wavelets que emprega filtros não lineares passa-baixas e passa-altas, baseados em aberturas e fechamentos, que por sua vez são construídos a partir dos operadores morfológicos clássicos de erosão e dilatação. Neste trabalho, são propostas aproximações aritméticas para esses dois operadores morfológicos clássicos, substituindo os desvios condicionais, presentes na Morfologia Matemática, por operações aritméticas de somas, subtrações e multiplicações, computacionalmente mais rápidas. O trabalho proposto compara o tempo estimado de execução entre as aproximações aritméticas propostas e as operações morfológicas clássicas utilizando a notação Big-Oh e também faz uso de estimativas baseadas em arquitetura de hardware pipeline. Em todas as estimativas e cenários reais, as aproximações morfológicas propostas são mais rápidas do que a morfologia clássica. Além disso, por não empregar unidade de hardware em ambiente pipeline para tratamento de desvios condicionais, as aproximações morfológicas propostas se tornam uma solução mais barata, ocupa menos espaço, mais propícia à miniaturização, consome menos energia e reduz o número de codificações da UC (Unidade de Controle). Logo, as Aproximações Morfológicas criadas são superiores à morfologia clássica nos principais requisitos para o bom funcionamento do hardware. Quanto à classificação, a Decomposição baseada em Aproximações Morfológicas alcança um desempenho médio de 84,65% na distinção entre casos normais, benignos e malignos. Os classificadores empregados são redes neurais ELM e SVM, cujas classes são definidas de acordo com os critérios da American College of Radiology. Foram usadas 355 imagens de mama adiposa da base de dados IRMA, com 233 casos normais, 66 benignos e 56 malignos. Como método de tratamento da base de dados, foram estudados pesos ponderando a fronteira de decisão das redes neurais. / According to the World Health Organization, breast cancer is the main cause of death of women round the world. From the clinical point of view, mammography is still the most effective diagnostic technology, given the wide diffusion of the use and interpretation of these images. According the state-of-the-art lesions classification on mammograms, wavelets have produced the best results from the viewpoint of precision, when used as a preprocessing step that decomposes the original image into approximation and detail images (vertical, horizontal and diagonal) in order to, from these components images, extract shape or texture attributes. This work proposes the decomposition Morphological-based in regions of interest on mammograms. The proposed method is inspired on wavelets decomposition employing nonlinear low-pass and high-pass filters, based on openings and closings, which are constructed from classical morphological operators of erosion and dilation. In this work, we propose approaches of classical morphology, replacing the conditional branches, present in Mathematical Morphology, by arithmetic operations of addition, subtraction and multiplication, computationally faster. The proposed work compares the estimated run time of proposed arithmetic approximations and classical morphological operations using Big-Oh notation and also the thesis uses notation based on pipeline hardware architecture. In all real scenarios, our morphological operations are faster than classical morphology. Also, by not employing hardware unit in pipeline environment for treating conditional branches, the proposed morphology approximation becomes a cheaper solution, occupies less space, more propitious to miniaturization, consumes less power, and reduces the Control Unit coding number. Then, our approaches of classical morphology are superior to classical morphology in key requirements of hardware solution. Regarding the classification, the proposed decomposition reaches an average performance of 84.65% in distinguishing normal, benign, and malignant cases. Classifiers are neural networks ELM and SVM, classes are defined according American College of Radiology criteria. They are employed 355 adipose breast images with 233 normal cases, 66 benign and 56 malignant. As database processing method, weights were studied considering the decision boundary of neural networks.

Les systèmes de vision artificielle sont de plus en plus présents en contexte routier. Ils sont installés sur l'infrastructure, pour la gestion du trafic, ou placés à l'intérieur du véhicule, pour proposer des aides à la conduite. Dans les deux cas, les systèmes de vision artificielle visent à augmenter la sécurité et à optimiser les déplacements. Une revue bibliographique retrace les origines et le développement des algorithmes de vision artificielle en contexte routier. Elle permet de démontrer l'importance des descripteurs d'images dans la chaîne de traitement des algorithmes. Elle se poursuit par une revue des descripteurs d'images avec une nouvelle approche source de nombreuses analyses, en les considérant en parallèle des applications finales. En conclusion, la revue bibliographique permet de déterminer quels sont les descripteurs d'images les plus représentatifs en contexte routier. Plusieurs bases de données contenant des images et les données météorologiques associées (ex : pluie, brouillard) sont ensuite présentées. Ces bases de données sont innovantes car l'acquisition des images et la mesure des conditions météorologiques sont effectuées en même temps et au même endroit. De plus, des capteurs météorologiques calibrés sont utilisés. Chaque base de données contient différentes scènes (ex: cible noir et blanc, piéton) et divers types de conditions météorologiques (ex: pluie, brouillard, jour, nuit). Les bases de données contiennent des conditions météorologiques naturelles, reproduites artificiellement et simulées numériquement. Sept descripteurs d'images parmi les plus représentatifs du contexte routier ont ensuite été sélectionnés et leur robustesse en conditions de pluie évaluée. Les descripteurs d'images basés sur l'intensité des pixels ou les contours verticaux sont sensibles à la pluie. A l'inverse, le descripteur de Harris et les descripteurs qui combinent différentes orientations sont robustes pour des intensités de pluie de 0 à 30 mm/h. La robustesse des descripteurs d'images en conditions de pluie diminue lorsque l'intensité de pluie augmente. Finalement, les descripteurs les plus sensibles à la pluie peuvent potentiellement être utilisés pour des applications de détection de la pluie par caméra.Le comportement d'un descripteur d'images en conditions météorologiques dégradées n'est pas forcément relié à celui de la fonction finale associée. Pour cela, deux détecteurs de piéton ont été évalués en conditions météorologiques dégradées (pluie, brouillard, jour, nuit). La nuit et le brouillard sont les conditions qui ont l'impact le plus important sur la détection des piétons. La méthodologie développée et la base de données associée peuvent être utilisées à nouveau pour évaluer d'autres fonctions finales (ex: détection de véhicule, détection de signalisation verticale).En contexte routier, connaitre les conditions météorologiques locales en temps réel est essentiel pour répondre aux deux enjeux que sont l'amélioration de la sécurité et l'optimisation des déplacements. Actuellement, le seul moyen de mesurer ces conditions le long des réseaux est l'installation de stations météorologiques. Ces stations sont coûteuses et nécessitent une maintenance particulière. Cependant, de nombreuses caméras sont déjà présentes sur le bord des routes. Une nouvelle méthode de détection des conditions météorologiques utilisant les caméras de surveillance du trafic est donc proposée. Cette méthode utilise des descripteurs d'images et un réseau de neurones. Elle répond à un ensemble de contraintes clairement établies afin de pouvoir détecter l'ensemble des conditions météorologiques en temps réel, mais aussi de pourvoir proposer plusieurs niveaux d'intensité. La méthode proposée permet de détecter les conditions normales de jour, de nuit, la pluie et le brouillard. Après plusieurs phases d'optimisation, la méthode proposée obtient de meilleurs résultats que ceux obtenus dans la littérature, pour des algorithmes comparables. / Computer vision systems are increasingly being used on roads. They can be installed along infrastructure for traffic monitoring purposes. When mounted in vehicles, they perform driver assistance functions. In both cases, computer vision systems enhance road safety and streamline travel.A literature review starts by retracing the introduction and rollout of computer vision algorithms in road environments, and goes on to demonstrate the importance of image descriptors in the processing chains implemented in such algorithms. It continues with a review of image descriptors from a novel approach, considering them in parallel with final applications, which opens up numerous analytical angles. Finally the literature review makes it possible to assess which descriptors are the most representative in road environments.Several databases containing images and associated meteorological data (e.g. rain, fog) are then presented. These databases are completely original because image acquisition and weather condition measurement are at the same location and the same time. Moreover, calibrated meteorological sensors are used. Each database contains different scenes (e.g. black and white target, pedestrian) and different kind of weather (i.e. rain, fog, daytime, night-time). Databases contain digitally simulated, artificial and natural weather conditions.Seven of the most representative image descriptors in road context are then selected and their robustness in rainy conditions is evaluated. Image descriptors based on pixel intensity and those that use vertical edges are sensitive to rainy conditions. Conversely, the Harris feature and features that combine different edge orientations remain robust for rainfall rates ranging in 0 – 30 mm/h. The robustness of image features in rainy conditions decreases as the rainfall rate increases. Finally, the image descriptors most sensitive to rain have potential for use in a camera-based rain classification application.The image descriptor behaviour in adverse weather conditions is not necessarily related to the associated final function one. Thus, two pedestrian detectors were assessed in degraded weather conditions (rain, fog, daytime, night-time). Night-time and fog are the conditions that have the greatest impact on pedestrian detection. The methodology developed and associated database could be reused to assess others final functions (e.g. vehicle detection, traffic sign detection).In road environments, real-time knowledge of local weather conditions is an essential prerequisite for addressing the twin challenges of enhancing road safety and streamlining travel. Currently, the only mean of quantifying weather conditions along a road network requires the installation of meteorological stations. Such stations are costly and must be maintained; however, large numbers of cameras are already installed on the roadside. A new method that uses road traffic cameras to detect weather conditions has therefore been proposed. This method uses a combination of a neural network and image descriptors applied to image patches. It addresses a clearly defined set of constraints relating to the ability to operate in real-time and to classify the full spectrum of meteorological conditions and grades them according to their intensity. The method differentiates between normal daytime, rain, fog and normal night-time weather conditions. After several optimisation steps, the proposed method obtains better results than the ones reported in the literature for comparable algorithms.

The relationship between post-mortem state and changes of biological tissue impedance has been investigated to serve as a basis for developing an in-vitro measurement method for monitoring the freshness of meat. The main challenges thereby are the reproducible measurement of the impedance of biological tissues and the classification method of their type and state. In order to realize reproducible tissue bio-impedance measurements, a suitable sensor taking into account the anisotropy of the biological tissue has been developed. It consists of cylindrical penetrating multi electrodes realizing good contacts between electrodes and the tissue. Experimental measurements have been carried out with different tissues and for a long period of time in order to monitor the state degradation with time. Measured results have been evaluated by means of the modified Fricke-Cole-Cole model. Results are reproducible and correspond to the expected behavior due to aging. An appropriate method for feature extraction and classification has been proposed using model parameters as features as input for classification using neural networks and fuzzy logic. A Multilayer Perceptron neural network (MLP) has been proposed for muscle type computing and the age computing and respectively freshness state of the meat. The designed neural network is able to generalize and to correctly classify new testing data with a high performance index of recognition. It reaches successful results of test equal to 100% for 972 created inputs for each muscle. An investigation of the influence of noise on the classification algorithm shows, that the MLP neural network has the ability to correctly classify the noisy testing inputs especially when the parameter noise is less than 0.6%. The success of classification is 100% for the muscles Longissimus Dorsi (LD) of beef, Semi-Membraneous (SM) of beef and Longissimus Dorsi (LD) of veal and 92.3% for the muscle Rectus Abdominis (RA) of veal. Fuzzy logic provides a successful alternative for easy classification. Using the Gaussian membership functions for the muscle type detection and trapezoidal member function for the classifiers related to the freshness detection, fuzzy logic realized an easy method of classification and generalizes correctly the inputs to the corresponding classes with a high level of recognition equal to 100% for meat type detection and with high accuracy for freshness computing equal to 84.62% for the muscle LD beef, 92.31 % for the muscle RA beef, 100 % for the muscle SM veal and 61.54% for the muscle LD veal. / Auf der Basis von Impedanzspektroskopie wurde ein neuartiges in-vitro-Messverfahren zur Überwachung der Frische von biologischem Gewebe entwickelt. Die wichtigsten Herausforderungen stellen dabei die Reproduzierbarkeit der Impedanzmessung und die Klassifizierung der Gewebeart sowie dessen Zustands dar. Für die Reproduzierbarkeit von Impedanzmessungen an biologischen Geweben, wurde ein zylindrischer Multielektrodensensor realisiert, der die 2D-Anisotropie des Gewebes berücksichtigt und einen guten Kontakt zum Gewebe realisiert. Experimentelle Untersuchungen wurden an verschiedenen Geweben über einen längeren Zeitraum durchgeführt und mittels eines modifizierten Fricke-Cole-Cole-Modells analysiert. Die Ergebnisse sind reproduzierbar und entsprechen dem physikalisch-basierten erwarteten Verhalten. Als Merkmale für die Klassifikation wurden die Modellparameter genutzt.

In this dissertation a method for the classification of emotion in speech recordings made in a customer service call centre of a large business is presented. The problem addressed here is that customer service analysts at large businesses have to listen to large numbers of call centre recordings in order to discover customer service-related issues. Since recordings where the customer exhibits emotion are more likely to contain useful information for service improvement than “neutral” ones, being able to identify those recordings should save a lot of time for the customer service analyst. MTN South Africa agreed to provide assistance for this project. The system that has been developed for this project can interface with MTN’s call centre database, download recordings, classify them according to their emotional content, and provide feedback to the user. The system faces the additional challenge that it is required to classify emotion notwith- standing the fact that the caller may have one of several South African accents. It should also be able to function with recordings made at telephone quality sample rates. The project identifies several speech features that can be used to classify a speech recording according to its emotional content. The project uses these features to research the general methods by which the problem of emotion classification in speech can be approached. The project examines both a K-Nearest Neighbours Approach and an Artificial Neural Network- Based Approach to classify the emotion of the speaker. Research is also done with regard to classifying a recording according to the gender of the speaker using a neural network approach. The reason for this classification is that the gender of a speaker may be useful input into an emotional classifier. The project furthermore examines the problem of identifying smaller segments of speech in a recording. In the typical call centre conversation, a recording may start with the agent greeting the customer, the customer stating his or her problem, the agent performing an action, during which time no speech occurs, the agent reporting back to the user and the call being terminated. The approach taken by this project allows the program to isolate these different segments of speech in a recording and discard segments of the recording where no speech occurs. This project suggests and implements a practical approach to the creation of a classifier in a commercial environment through its use of a scripting language interpreter that can train a classifier in one script and use the trained classifier in another script to classify unknown recordings. The project also examines the practical issues involved in implementing an emotional clas- sifier. It addresses the downloading of recordings from the call centre, classifying the recording and presenting the results to the customer service analyst. AFRIKAANS : n Metode vir die klassifisering van emosie in spraakopnames in die oproepsentrum van ’n groot sake-onderneming word in hierdie verhandeling aangebied. Die probleem wat hierdeur aangespreek word, is dat kli¨entediens ontleders in ondernemings na groot hoeveelhede oproepsentrum opnames moet luister ten einde kli¨entediens aangeleenthede te identifiseer. Aangesien opnames waarin die kli¨ent emosie toon, heel waarskynlik nuttige inligting bevat oor diensverbetering, behoort die vermo¨e om daardie opnames te identifiseer vir die analis baie tyd te spaar. MTN Suid-Afrika het ingestem om bystand vir die projek te verleen. Die stelsel wat ontwikkel is kan opnames vanuit MTN se oproepsentrum databasis verkry, klassifiseer volgens emosionele inhoud en terugvoering aan die gebruiker verskaf. Die stelsel moet die verdere uitdaging kan oorkom om emosie te kan klassifiseer nieteenstaande die feit dat die spreker een van verskeie Suid-Afrikaanse aksente het. Dit moet ook in staat wees om opnames wat gemaak is teen telefoon gehalte tempos te analiseer. Die projek identifiseer verskeie spraak eienskappe wat gebruik kan word om ’n opname volgens emosionele inhoud te klassifiseer. Die projek gebruik hierdie eienskappe om die algemene metodes waarmee die probleem van emosie klassifisering in spraak benader kan word, na te vors. Die projek gebruik ’n K-Naaste Bure en ’n Neurale Netwerk benadering om die emosie van die spreker te klassifiseer. Navorsing is voorts gedoen met betrekking tot die klassifisering van die geslag van die spreker deur ’n neurale netwerk. Die rede vir hierdie klassifisering is dat die geslag van die spreker ’n nuttige inset vir ’n emosie klassifiseerder mag wees. Die projek ondersoek ook die probleem van identifisering van spraakgedeeltes in ’n opname. In ’n tipiese oproepsentrum gesprek mag die opname begin met die agent wat die kli¨ent groet, die kli¨ent wat sy of haar probleem stel, die agent wat ’n aksie uitvoer sonder spraak, die agent wat terugrapporteer aan die gebruiker en die oproep wat be¨eindig word. Die benadering van hierdie projek laat die program toe om hierdie verskillende gedeeltes te isoleer uit die opname en om gedeeltes waar daar geen spraak plaasvind nie, uit te sny. Die projek stel ’n praktiese benadering vir die ontwikkeling van ’n klassifiseerder in ’n kommersi¨ele omgewing voor en implementeer dit deur gebruik te maak van ’n programeer taal interpreteerder wat ’n klassifiseerder kan oplei in een program en die opgeleide klassifiseerder gebruik om ’n onbekende opname te klassifiseer met behulp van ’n ander program. Die projek ondersoek ook die praktiese aspekte van die implementering van ’n emosionele klassifiseerder. Dit spreek die aflaai van opnames uit die oproep sentrum, die klassifisering daarvan, en die aanbieding van die resultate aan die kli¨entediens analis, aan. Copyright / Dissertation (MEng)--University of Pretoria, 2010. / Electrical, Electronic and Computer Engineering / unrestricted

With the introduction of new input devices, such as multi-touch surface displays, the Nintendo WiiMote, the Microsoft Kinect, and the Leap Motion sensor, among others, the field of Human-Computer Interaction (HCI) finds itself at an important crossroads that requires solving new challenges. Given the amount of three-dimensional (3D) data available today, 3D navigation plays an important role in 3D User Interfaces (3DUI). This dissertation deals with multi-touch, 3D navigation, and how users can explore 3D virtual worlds using a multi-touch, non-stereo, desktop display. The contributions of this dissertation include a feature-extraction algorithm for multi-touch displays (FETOUCH), a multi-touch and gyroscope interaction technique (GyroTouch), a theoretical model for multi-touch interaction using high-level Petri Nets (PeNTa), an algorithm to resolve ambiguities in the multi-touch gesture classification process (Yield), a proposed technique for navigational experiments (FaNS), a proposed gesture (Hold-and-Roll), and an experiment prototype for 3D navigation (3DNav). The verification experiment for 3DNav was conducted with 30 human-subjects of both genders. The experiment used the 3DNav prototype to present a pseudo-universe, where each user was required to find five objects using the multi-touch display and five objects using a game controller (GamePad). For the multi-touch display, 3DNav used a commercial library called GestureWorks in conjunction with Yield to resolve the ambiguity posed by the multiplicity of gestures reported by the initial classification. The experiment compared both devices. The task completion time with multi-touch was slightly shorter, but the difference was not statistically significant. The design of experiment also included an equation that determined the level of video game console expertise of the subjects, which was used to break down users into two groups: casual users and experienced users. The study found that experienced gamers performed significantly faster with the GamePad than casual users. When looking at the groups separately, casual gamers performed significantly better using the multi-touch display, compared to the GamePad. Additional results are found in this dissertation.

Advances in unsupervised learning and deep neural networks have led to increased performance in a number of domains, and to the ability to draw strong comparisons between the biological method of self-organization conducted by the brain and computational mechanisms. This thesis aims to use real-world data to tackle two areas in the domain of computer vision which have biological equivalents: feature detection and motion tracking. The aforementioned advances have allowed efficient learning of feature representations directly from large sets of unlabeled data instead of using traditional handcrafted features. The first part of this thesis evaluates such representations by comparing regularization and preprocessing methods which incorporate local neighbouring information during training on a single-layer neural network. The networks are trained and tested on the Hollywood2 video dataset, as well as the static CIFAR-10, STL-10, COIL-100, and MNIST image datasets. The induction of topography or simple image blurring via Gaussian filters during training produces better discriminative features as evidenced by the consistent and notable increase in classification results that they produce. In the visual domain, invariant features are desirable such that objects can be classified despite transformations. It is found that most of the compared methods produce more invariant features, however, classification accuracy does not correlate to invariance. The second, and paramount, contribution of this thesis is a biologically-inspired model to explain the emergence of motion tracking behaviour in early development using unsupervised learning. The model’s self-organization is biased by an original concept called retinal constancy, which measures how similar visual contents are between successive frames. In the proposed two-layer deep network, when exposed to real-world video, the first layer learns to encode visual motion, and the second layer learns to relate that motion to gaze movements, which it perceives and creates through bi-directional nodes. This is unique because it uses general machine learning algorithms, and their inherent generative properties, to learn from real-world data. It also implements a biological theory and learns in a fully unsupervised manner. An analysis of its parameters and limitations is conducted, and its tracking performance is evaluated. Results show that this model is able to successfully follow targets in real-world video, despite being trained without supervision on real-world video.

Analysing handwritten texts and creating labelled data sets can facilitate novel research on languages and advanced computerized analysis of authors works. However, few handwritten works have word wise labelling or data sets associated with them. More often a transcription of the text is available, but without any exact coupling between words in the transcript and word representations in the document images. Can an algorithm be created that will take only an image of handwritten text and a corresponding transcript and return a partial alignment and data set? An algorithm is developed in this thesis that explores the use of a convolutional neural network trained on English handwritten text to be able to align some words on pages and create a data set given a handwritten page image and a transcript. This algorithm is tested on handwritten English text. The algorithm is also tested on Swedish shorthand, which was the inspiration for the development of the algorithm in this work. In testing on several pages of handwritten English text, the algorithm reaches an overall average classification of 68% of words on one page with 0% miss-classification of those words. On a sequence of pages, the algorithm reaches 84% correctly classified words on 10 pages and produces a data set of 551 correctly labelled word images. This after being shown 10 pages with an average of 70.6 words on each page, with0% miss-classification. / Analys av handskrivna texter och skapande av dataset kan främja ny forskning inom språk och avancerad datoranalys av olika författares verk. Det finns dock få handskrivna verk med information om vad varje handskrivet ord betecknar eller dataset relaterade till texten. Oftare finns en transkribering av texten, utan någon exakt koppling mellan de transkriberade orden och handskrivna ord i bilden av ett dokument. Genom att skapa en algoritm som kan ta tillvara handskrivna texter och motsvarande transkription kan potentiellt fler verk datoranalyseras. Kan en algoritm skapas som bara tar in en bild av ett handskrivet dokument och en motsvarande transkription och som returnerar en partiell placering av ord till ordbilder och ett dataset? En algoritm skapas i detta arbete som utforskar möjligheten att använda ett djupt neuralt nätverk tränat på engelsk handskriven text för att koppla ord i ett dokumentet till en transkription, och använda dessa för att skapa ett dataset. Denna algoritm är testad på engelsk handskriven text. Algoritmen testas också på svensk stenografi, vilket är inspirationen till skapandet av algoritmen. Algoritmen testades på ett antal sidor handskriven engelsk text. Där kunde algoritmen klassificera i genomsnitt 68% av orden på en handskriven sida med 0% av dessa ord felklassificerade. På en serie sidor når algoritmen en genomsnittlig klassificering av 84% klassificerade ord, och producerar ett dataset av 551 korrekt klassificerade ordbilder. Detta är efter att ha visat algoritmen 10 sidor med i snitt 70.6 ord per sida. I dessa test nåddes också en felklassificering på 0%.