Neste trabalho foram propostas técnicas para a segmentação automática de mamogramas e para a detecção de assimetrias entre mamogramas esquerdo e direito. A segmentação é realizada através de três técnicas computacionais para a identificação de três importantes regiões anatômicas nos mamogramas: borda da mama, músculo peitoral e disco fibro-glandular. O primeiro método focaliza a identificação da borda da mama através do uso de um modelo de contorno ativo especialmente projetado para esse propósito. Neste estágio, a borda da mama é automaticamente demarcada, todos os artefatos fora dessa região são eliminados, e a região de interesse usada para a detecção do músculo peitoral é definida. No próximo estágio, a borda do músculo peitoral é determinada usando uma técnica multiresolução baseada na representação Gabor wavelets. Finalmente, um modelo de densidades da mama, baseado no modelo da mistura finita de Gaussianas, é proposto para a representação de quatro categorias de tecidos mamários com diferentes densidades. O disco fibro-glandular é identificado através da aplicação de um limiar sob as classes de densidades determinadas no modelo. Os métodos propostos foram aplicados em 84 imagens de mamogramas de projeções médio-laterais oblíqüas da base de dados Mini-MIAS (\"Mammographic Image Analysis Society\", London, UK). A avaliação dos resultados dos procedimentos de segmentação da borda da mama e borda do músculo peitoral foi realizada com base no percentual de pixels falso-positivos (FPs) e falso-negativos (FNs) determinados por comparação entre os contornos verdadeiros e os contornos automaticamente identificados. As taxas médias de FPs e FNs para as bordas da mama e do músculo peitoral foram, respectivamente, de 0,41% e 0,58%, e 1,78% e 5,77%. A segmentação dos discos fibro-glandulares foi subjetivamente classificada por radiologistas e os resultados indicaram que em mais de 80% dos casos a segmentação foi ) considerada aceitável para o uso em sistemas de auxílio ao diagnóstico. A detecção de assimetrias foi realizada usando informações direcionais, obtidas a partir da representação multiresolução Gabor wavelets, e de informações de forma e densidade, extraídas dos discos fibro-glandulares dos mamogramas esquerdo e direito. No procedimento de análise direcional, uma representação wavelet formada por filtros de Gabor bidimensionais com variação em freqüência e orientação, especialmente projetadas para reduzir a redundância na representação, é aplicada para uma dada imagem. As respostas dos filtros para diferentes escalas e orientações são analisadas através da transformada de Karhunen-Loève (KL) e pelo método de limiarização de Otsu. A transformada KL é aplicada para selecionar os componentes principais das respostas dos filtros, preservando apenas os elementos direcionais mais relevantes que aparecem em todas as escalas. Os componentes principais selecionados e limiarizados pela técnica de Otsu são usados para obter as imagens de magnitude e fase dos componentes direcionais da imagem. Medidas estatísticas extraídas dos diagramas de rosa calculados a partir das imagens de fase são usadas para a análise quantitativa e qualitativa dos padrões orientados. Um total de 11 atributos é extraído dos discos fibro-glandulares segmentados dos mamogramas esquerdo e direito, e a diferença calculada para cada par de atributos é usada como uma medida para a detecção de assimetrias. Um total de 88 imagens (22 casos normais, 14 casos de densidades assimétricas e 8 casos de distorções de arquitetura) da base de dados Mini-MIAS foram usadas para avaliar o método proposto. A combinação exaustiva dos atributos juntamente com a análise de componentes principais foi usada para selecionar o melhor subgrupo de atributos. A classificação foi realizada através de classificadores de Bayes (linear e quadrático) ) e usando o método \"leave-one-out\". Uma taxa de classificação correta de 84,44% foi alcançada. / In this work, techniques are proposed for the automatic segmentation of mammograms and detection of asymmetries between left and right mammograms. The segmentation is performed by using three computational techniques for the identification of three important anatomical regions of mammograms: the skin-air boundary, the pectoral muscle, and the fibro-glandular disc. The first method focuses on the identification of the skin-air boundary by using an active contour model algorithm specially tailored for this purpose. In this stage, the skin-air boundary is demarcated, all artefacts outside the breast region are eliminated, and the region of interest for detection of the pectoral muscle is defined. In the next stage, the edge of the pectoral muscle is determined by using a multiresolution technique based upon a Gabor wavelets representation. Finally, a density breast model based upon a Gaussian mixture model is proposed for the representation of four categories of different density tissues in the breast. The fibro-glandular disc is identified by thresholding the density categories of the model. The methods proposed were applied to 84 images of medio-lateral oblique mammograms from the Mini-MIAS (Mammographic Image Analysis Society, London, U.K.) database. The evaluation of the skin-air boundary and the pectoral muscle edge were performed based upon the percentage of false-positive (FP) and false-negative (FN) pixels determined by comparison between the true contours and the contours automatically identified. The FP and FN average rates for the skin-air boundary and the pectoral muscle edge were, respectively, 0.41% and 0.58%, and 1.78% and 5.77%. Two radiologists subjectively rated the segmentation of the fibro-glandular disc and the results indicate that in more than 80% of the cases, the segmentation was considered acceptable for a Computer Aided Diagnosis purposes. Detection of asymmetries (continua) (continuação) is performed by using directional information, obtained from a multiresolution Gabor wavelets representation, and shape and density information, extracted from the fibro-glandular discs of left and right mammograms. In the directional procedure, a particular wavelet scheme with 2-D Gabor filters as elementary functions with varying tuning frequency and orientation, specifically designed in order to reduce the redundancy in the wavelet-based representation, is applied to the given image. The filter responses for different scales and orientation are analyzed by using the Karhunen-Loève (KL) transform and Otsu\'s method of thresholding. The KL transform is applied to select the principal components of the filter responses, preserving only the most relevant directional elements appearing at all scales. The selected principal components are thresholded by using Otsu\'s method and used to obtain the magnitude and phase of the image directional components. Rose diagrams computed from the phase images and statistical measures computed thereof are used for quantitative and qualitative analysis of the oriented patterns. A total of 11 features are also extracted from the segmented fibro-glandular discs of left-right mammograms, and the difference of each feature pair is used as a measure for detecting asymmetries. A total of 88 images from 22 normal cases, 14 asymmetric cases, and 8 architectural distortion cases from the Mini-MIAS database were used to evaluate the scheme. An exhaustive combination of the features along with the principal components analysis was used to select the best feature set. The classification was performed by using two Bayes\' classifiers (linear and quadratic) and the leave-one-out methodology. Average classification accuracy up to 84.44% was achieved.

Neste trabalho foram propostas técnicas para a segmentação automática de mamogramas e para a detecção de assimetrias entre mamogramas esquerdo e direito. A segmentação é realizada através de três técnicas computacionais para a identificação de três importantes regiões anatômicas nos mamogramas: borda da mama, músculo peitoral e disco fibro-glandular. O primeiro método focaliza a identificação da borda da mama através do uso de um modelo de contorno ativo especialmente projetado para esse propósito. Neste estágio, a borda da mama é automaticamente demarcada, todos os artefatos fora dessa região são eliminados, e a região de interesse usada para a detecção do músculo peitoral é definida. No próximo estágio, a borda do músculo peitoral é determinada usando uma técnica multiresolução baseada na representação Gabor wavelets. Finalmente, um modelo de densidades da mama, baseado no modelo da mistura finita de Gaussianas, é proposto para a representação de quatro categorias de tecidos mamários com diferentes densidades. O disco fibro-glandular é identificado através da aplicação de um limiar sob as classes de densidades determinadas no modelo. Os métodos propostos foram aplicados em 84 imagens de mamogramas de projeções médio-laterais oblíqüas da base de dados Mini-MIAS (\"Mammographic Image Analysis Society\", London, UK). A avaliação dos resultados dos procedimentos de segmentação da borda da mama e borda do músculo peitoral foi realizada com base no percentual de pixels falso-positivos (FPs) e falso-negativos (FNs) determinados por comparação entre os contornos verdadeiros e os contornos automaticamente identificados. As taxas médias de FPs e FNs para as bordas da mama e do músculo peitoral foram, respectivamente, de 0,41% e 0,58%, e 1,78% e 5,77%. A segmentação dos discos fibro-glandulares foi subjetivamente classificada por radiologistas e os resultados indicaram que em mais de 80% dos casos a segmentação foi ) considerada aceitável para o uso em sistemas de auxílio ao diagnóstico. A detecção de assimetrias foi realizada usando informações direcionais, obtidas a partir da representação multiresolução Gabor wavelets, e de informações de forma e densidade, extraídas dos discos fibro-glandulares dos mamogramas esquerdo e direito. No procedimento de análise direcional, uma representação wavelet formada por filtros de Gabor bidimensionais com variação em freqüência e orientação, especialmente projetadas para reduzir a redundância na representação, é aplicada para uma dada imagem. As respostas dos filtros para diferentes escalas e orientações são analisadas através da transformada de Karhunen-Loève (KL) e pelo método de limiarização de Otsu. A transformada KL é aplicada para selecionar os componentes principais das respostas dos filtros, preservando apenas os elementos direcionais mais relevantes que aparecem em todas as escalas. Os componentes principais selecionados e limiarizados pela técnica de Otsu são usados para obter as imagens de magnitude e fase dos componentes direcionais da imagem. Medidas estatísticas extraídas dos diagramas de rosa calculados a partir das imagens de fase são usadas para a análise quantitativa e qualitativa dos padrões orientados. Um total de 11 atributos é extraído dos discos fibro-glandulares segmentados dos mamogramas esquerdo e direito, e a diferença calculada para cada par de atributos é usada como uma medida para a detecção de assimetrias. Um total de 88 imagens (22 casos normais, 14 casos de densidades assimétricas e 8 casos de distorções de arquitetura) da base de dados Mini-MIAS foram usadas para avaliar o método proposto. A combinação exaustiva dos atributos juntamente com a análise de componentes principais foi usada para selecionar o melhor subgrupo de atributos. A classificação foi realizada através de classificadores de Bayes (linear e quadrático) ) e usando o método \"leave-one-out\". Uma taxa de classificação correta de 84,44% foi alcançada. / In this work, techniques are proposed for the automatic segmentation of mammograms and detection of asymmetries between left and right mammograms. The segmentation is performed by using three computational techniques for the identification of three important anatomical regions of mammograms: the skin-air boundary, the pectoral muscle, and the fibro-glandular disc. The first method focuses on the identification of the skin-air boundary by using an active contour model algorithm specially tailored for this purpose. In this stage, the skin-air boundary is demarcated, all artefacts outside the breast region are eliminated, and the region of interest for detection of the pectoral muscle is defined. In the next stage, the edge of the pectoral muscle is determined by using a multiresolution technique based upon a Gabor wavelets representation. Finally, a density breast model based upon a Gaussian mixture model is proposed for the representation of four categories of different density tissues in the breast. The fibro-glandular disc is identified by thresholding the density categories of the model. The methods proposed were applied to 84 images of medio-lateral oblique mammograms from the Mini-MIAS (Mammographic Image Analysis Society, London, U.K.) database. The evaluation of the skin-air boundary and the pectoral muscle edge were performed based upon the percentage of false-positive (FP) and false-negative (FN) pixels determined by comparison between the true contours and the contours automatically identified. The FP and FN average rates for the skin-air boundary and the pectoral muscle edge were, respectively, 0.41% and 0.58%, and 1.78% and 5.77%. Two radiologists subjectively rated the segmentation of the fibro-glandular disc and the results indicate that in more than 80% of the cases, the segmentation was considered acceptable for a Computer Aided Diagnosis purposes. Detection of asymmetries (continua) (continuação) is performed by using directional information, obtained from a multiresolution Gabor wavelets representation, and shape and density information, extracted from the fibro-glandular discs of left and right mammograms. In the directional procedure, a particular wavelet scheme with 2-D Gabor filters as elementary functions with varying tuning frequency and orientation, specifically designed in order to reduce the redundancy in the wavelet-based representation, is applied to the given image. The filter responses for different scales and orientation are analyzed by using the Karhunen-Loève (KL) transform and Otsu\'s method of thresholding. The KL transform is applied to select the principal components of the filter responses, preserving only the most relevant directional elements appearing at all scales. The selected principal components are thresholded by using Otsu\'s method and used to obtain the magnitude and phase of the image directional components. Rose diagrams computed from the phase images and statistical measures computed thereof are used for quantitative and qualitative analysis of the oriented patterns. A total of 11 features are also extracted from the segmented fibro-glandular discs of left-right mammograms, and the difference of each feature pair is used as a measure for detecting asymmetries. A total of 88 images from 22 normal cases, 14 asymmetric cases, and 8 architectural distortion cases from the Mini-MIAS database were used to evaluate the scheme. An exhaustive combination of the features along with the principal components analysis was used to select the best feature set. The classification was performed by using two Bayes\' classifiers (linear and quadratic) and the leave-one-out methodology. Average classification accuracy up to 84.44% was achieved.

The problem of characterization of active sonar target response has important applications in many fields, including the currently cost-prohibitive recovery of unexploded ordinance on the ocean floor. We present a method for recognizing these objects using a multidisciplinary approach that fuses machine learning, signal processing, and feature engineering. In short, by taking inspiration from other fields, we solve the problem of object recognition in shallow water in an inexpensive way. These techniques add to the body of explored knowledge in the field of active sonar processing and address real-world problems in the process.

With an increasing emphasis on personal privacy, security, and convenience, the security of identification system is an important issue nowadays. In this thesis, two intelligent identification systems, laser speckle image identification system and laser-based finger biometric system, are proposed to perform superior solutions for identification applications. In laser speckle image identification system, we investigated the characteristics of laser speckle as well as proposed an appropriate algorithm to establish this system. The proposed algorithm is a coarse-to-fine process which identifies laser speckle images systematically. In laser-based finger biometric system, a new biometric approach is described to proceed personal identification using a scanner with a low power laser scans across the surface of the finger and continuously recording the reflected intensity at a fixed position. Experimental results show that the recognition rates of the proposed system are both 100%.

Iris recognition constitutes one of the most powerful method for the iden- tification and authentication of people today. This thesis aims to describe the algorithms used in a sophisticated and mathematically correct way, while re- maining comprehensible. The description of these algorithms is not the only objective of this thesis; the reason they were chosen and potential improvements or substitutions are also discussed. The background of iris recognition, its use in cryptosystems, and the application of error-correcting codes are investigated as well.

Iris recognition constitutes one of the most powerful method for the iden- tification and authentication of people today. This thesis aims to describe the algorithms used by a mathematical apparatus. The description of these algo- rithms is not the only objective of this thesis; the reason they were chosen and potential improvements or substitutions are also discussed. The background of iris recognition, its use in cryptosystems, and the application of error-correcting codes are investigated as well. The second version of the thesis eliminates errata and a quantum of inaccu- racies discovered in the first version, especially in the ROI Definition, the Hough Transform and the Feature Extraction sections. Besides that, it also contains se- veral new propositions. Last, but not least, it shows a potential implementation of the algorithms described by appending pseudocodes to the relevant sections. 1

CoordenaÃÃo de AperfeiÃoamento de NÃvel Superior / In this dissertation, the face recognition problem is investigated from the standpoint of multilinear algebra, more specifically the tensor decomposition, and by making use of Gabor wavelets. The feature extraction occurs in two stages: first the Gabor wavelets are applied holistically in feature selection; Secondly facial images are modeled as a higher-order tensor according to the multimodal factors present. Then, the HOSVD is applied to separate the multimodal factors of the images. The proposed facial recognition approach exhibits higher average success rate and stability when there is variation in the various multimodal factors such as facial position, lighting condition and facial expression. We also propose a systematic way to perform cross-validation on tensor models to estimate the error rate in face recognition systems that explore the nature of the multimodal ensemble. Through the random partitioning of data organized as a tensor, the mode-n cross-validation provides folds as subtensors extracted of the desired mode, featuring a stratified method and susceptible to repetition of cross-validation with different partitioning. / Nesta dissertaÃÃo o problema de reconhecimento facial à investigado do ponto de vista da Ãlgebra multilinear, mais especificamente por meio de decomposiÃÃes tensoriais fazendo uso das wavelets de Gabor. A extraÃÃo de caracterÃsticas ocorre em dois estÃgios: primeiramente as wavelets de Gabor sÃo aplicadas de maneira holÃstica na seleÃÃo de caracterÃsticas; em segundo as imagens faciais sÃo modeladas como um tensor de ordem superior de acordo com o fatores multimodais presentes. Com isso aplicamos a decomposiÃÃo tensorial Higher Order Singular Value Decomposition (HOSVD) para separar os fatores que influenciam na formaÃÃo das imagens. O mÃtodo de reconhecimento facial proposto possui uma alta taxa de acerto e estabilidade quando hà variaÃÃo nos diversos fatores multimodais, tais como, posiÃÃo facial, condiÃÃo de iluminaÃÃo e expressÃo facial. Propomos ainda uma maneira sistemÃtica para realizaÃÃo da validaÃÃo cruzada em modelos tensoriais para estimaÃÃo da taxa de erro em sistemas de reconhecimento facial que exploram a natureza multilinear do conjunto de imagens. AtravÃs do particionamento aleatÃrio dos dados organizado como um tensor, a validaÃÃo cruzada modo-n proporciona a criaÃÃo de folds extraindo subtensores no modo desejado, caracterizando um mÃtodo estratificado e susceptÃvel a repetiÃÃes da validaÃÃo cruzada com diferentes particionamentos.

Biometric identification systems have several distinct advantages over other authentication technologies, such as passwords, in reliably recognising individuals. Iris based recognition is one such biometric recognition system. Unlike other biometrics such as fingerprints or face images, the distinct aspect of the iris comes from its randomly distributed features. The patterns of these randomly distributed features on the iris have been proved to be fixed in a person's lifetime, and are stable over time for healthy eyes except for the distortions caused by the constriction and dilation of the pupil. The distortion of the iris pattern caused by pupillary activity, which is mainly due changes in ambient lighting conditions, can be significant. One important question that arises from this is: How closely do two different iris images of the same person, taken at different times using different cameras, in different environments, and under different lighting conditions, agree with each other? It is also problematic for iris recognition systems to correctly identify a person when his/her pupil size is very different from the person's iris images, used at the time of constructing the system's data-base. To date, researchers in the field of iris recognition have made attempts to address this problem, with varying degrees of success. However, there is still a need to conduct in-depth investigations into this matter in order to arrive at more reliable solutions. It is therefore necessary to study the behaviour of iris surface deformation caused by the change of lighting conditions. In this thesis, a study of the physiological behaviour of pupil size variation under different normal indoor lighting conditions (100 lux ~ 1,200 lux) and brightness levels is presented. The thesis also presents the results of applying Elastic Graph Matching (EGM) tracking techniques to study the mechanisms of iris surface deformation. A study of the pupil size variation under different normal indoor lighting conditions was conducted. The study showed that the behaviour of the pupil size can be significantly different from one person to another under the same lighting conditions. There was no evidence from this study to show that the exact pupil sizes of an individual can be determined at a given illumination level. However, the range of pupil sizes can be estimated for a range of specific lighting conditions. The range of average pupil sizes under normal indoor lighting found was between 3 mm and 4 mm. One of the advantages of using EGM for iris surface deformation tracking is that it incorporates the benefit of the use of Gabor wavelets to encode the iris features for tracking. The tracking results showed that the radial stretch of the iris surface is nonlinear. However, the amount of extension of iris surface at any point on the iris during the stretch is approximately linear. The analyses of the tracking results also showed that the behaviour of iris surface deformation is different from one person to another. This implies that a generalised iris surface deformation model cannot be established for personal identification. However, a deformation model can be established for every individual based on their analysis result, which can be useful for personal verification using the iris. Therefore, analysis of the tracking results of each individual was used to model iris surface deformations for that individual. The model was able to estimate the movement of a point on the iris surface at a particular pupil size. This makes it possible to estimate and construct the 2D deformed iris image of a desired pupil size from a given iris image of another different pupil size. The estimated deformed iris images were compared with their actual images for similarity, using an intensitybased (zero mean normalised cross-correlation). The result shows that 86% of the comparisons have over 65% similarity between the estimated and actual iris image. Preliminary tests of the estimated deformed iris images using an open-source iris recognition algorithm have showed an improved personal verification performance. The studies presented in this thesis were conducted using a very small sample of iris images and therefore should not be generalised, before further investigations are conducted.

The thesis deals with image classifiers and their implementation using FPGA technology. There are discussed weak and strong classifiers in the work. As an example of strong classifiers, the AdaBoost algorithm is described. In the case of weak classifiers, basic types of feature classifiers are shown, including Haar and Gabor wavelets. The rest of work is primarily focused on LBP, LRP and LR classifiers, which are well suitable for efficient implementation in FPGAs. With these classifiers is designed pseudo-parallel architecture. Process of classifications is divided on software and hardware parts. The thesis deals with hardware part of classifications. The designed classifier is very fast and produces results of classification every clock cycle.

Αντικείμενο της παρούσας εργασία είναι ο υπολογισμός και η ταξινόμηση, με βάση το σύστημα, BIRADS της πυκνότητας του μαστού από εικόνες μαστογραφίας. Στα πλαίσια της προσπάθειας αυτής αναπτύχθηκε ολοκληρωμένο υπολογιστικό σύστημα σε γραφικό περιβάλλον ως λογισμικό πακέτο, σε γλώσσα Visual C++ .NET . Το υπολογιστικό αυτό σύστημα δέχεται σαν είσοδο εικόνες μαστογραφίας σε οποιοδήποτε από τα δημοφιλή bitmap format εικόνων όπως jpeg και tiff καθώς και DICOM αρχεία. Η λειτουργία του μπορεί να χωριστεί σε τρία στάδια: το στάδιο της προεπεξεργασίας, το στάδιο απομόνωσης της περιοχής του μαστού και το στάδιο καθορισμού της πυκνότητας του μαστού. Στο πρώτο στάδιο παρέχονται μια σειρά από στοιχειώδη εργαλεία επεξεργασίας εικόνας όπως εργαλεία περιστροφής, αποκοπής και αλλαγής αντίθεσης . Επιπρόσθετα παρέχεται η δυνατότητα Ανισοτροπικού Φιλτραρίσματος της εικόνας. Στο δεύτερο στάδιο γίνεται η απομόνωση της περιοχής του μαστού είτε απευθείας από τον χρήστη είτε αυτόματα με χρήση των ιδιοτήτων του μονογονικού (monogenic) σήματος για την αφαίρεση του παρασκηνίου (background) καθώς και κυματιδίων Gabor για τον διαχωρισμού του θωρακικού μυός. Στο τρίτο στάδιο παρέχεται η δυνατότητα ταξινόμησης της πυκνότητας του μαστού από τον χρήστη με τον καθορισμό κατάλληλου κατωφλίου των επιπέδων γκρίζου της εικόνας αλλά και η δυνατότητα αυτόματης ταξινόμησης της πυκνότητας του μαστού κατά BIRADS με χρήση Δομικών Στοιχείων Υφής (textons) και της τεχνικής pLSA. Όλες οι παραπάνω λειτουργίες παρέχονται μέσω μίας κατά το δυνατόν φιλικότερης προς τον χρήστη διεπαφής. / The present thesis aims at the classification of breast tissue according to BIRADS system based on texture features. To this end an integrated software system was developed in visual C ++. The system takes as inputs pictures in most of the popular bitmap formats like .jpeg and .till as well as DICOM. The functionality of the system is provided by three modules: (a) pre-processing module, (b) breast segmentation module and (c) the breast tissue density classification module. In the pre-processing module a set tools for image manipulation (rotation, crop, gray level adjustment) are available which are accompanied by the ability to perform anisotropic filtering to the input image. In the second module, the user has the ability to interactively define the actual borders of the breast or ask the system to perform it automatically. Automatic segmentation is a two step procedure; in the first step breast tissue is separated from its background by using the characteristics of monogenic signals, while in the second step the pectoral muscle region is subtracted using Gabor wavelets. In the density classification module the user can either ask for a calculation of breast density based on user-defined grey level threshold or perform an automatic BIRADS-based classification using texture characteristics in conjunction with Probabilistic Latent Semantic Analysis (pLSA) algorithm. Special emphasis was given to the development of a functional and user-friendly interface.