abstract: Texture analysis plays an important role in applications like automated pattern inspection, image and video compression, content-based image retrieval, remote-sensing, medical imaging and document processing, to name a few. Texture Structure Analysis is the process of studying the structure present in the textures. This structure can be expressed in terms of perceived regularity. Our human visual system (HVS) uses the perceived regularity as one of the important pre-attentive cues in low-level image understanding. Similar to the HVS, image processing and computer vision systems can make fast and efficient decisions if they can quantify this regularity automatically. In this work, the problem of quantifying the degree of perceived regularity when looking at an arbitrary texture is introduced and addressed. One key contribution of this work is in proposing an objective no-reference perceptual texture regularity metric based on visual saliency. Other key contributions include an adaptive texture synthesis method based on texture regularity, and a low-complexity reduced-reference visual quality metric for assessing the quality of synthesized textures. In order to use the best performing visual attention model on textures, the performance of the most popular visual attention models to predict the visual saliency on textures is evaluated. Since there is no publicly available database with ground-truth saliency maps on images with exclusive texture content, a new eye-tracking database is systematically built. Using the Visual Saliency Map (VSM) generated by the best visual attention model, the proposed texture regularity metric is computed. The proposed metric is based on the observation that VSM characteristics differ between textures of differing regularity. The proposed texture regularity metric is based on two texture regularity scores, namely a textural similarity score and a spatial distribution score. In order to evaluate the performance of the proposed regularity metric, a texture regularity database called RegTEX, is built as a part of this work. It is shown through subjective testing that the proposed metric has a strong correlation with the Mean Opinion Score (MOS) for the perceived regularity of textures. The proposed method is also shown to be robust to geometric and photometric transformations and outperforms some of the popular texture regularity metrics in predicting the perceived regularity. The impact of the proposed metric to improve the performance of many image-processing applications is also presented. The influence of the perceived texture regularity on the perceptual quality of synthesized textures is demonstrated through building a synthesized textures database named SynTEX. It is shown through subjective testing that textures with different degrees of perceived regularities exhibit different degrees of vulnerability to artifacts resulting from different texture synthesis approaches. This work also proposes an algorithm for adaptively selecting the appropriate texture synthesis method based on the perceived regularity of the original texture. A reduced-reference texture quality metric for texture synthesis is also proposed as part of this work. The metric is based on the change in perceived regularity and the change in perceived granularity between the original and the synthesized textures. The perceived granularity is quantified through a new granularity metric that is proposed in this work. It is shown through subjective testing that the proposed quality metric, using just 2 parameters, has a strong correlation with the MOS for the fidelity of synthesized textures and outperforms the state-of-the-art full-reference quality metrics on 3 different texture databases. Finally, the ability of the proposed regularity metric in predicting the perceived degradation of textures due to compression and blur artifacts is also established. / Dissertation/Thesis / Ph.D. Electrical Engineering 2014

Este projeto descreve o desenvolvimento, estudo e aplicação de descritores fractais em análise de texturas. Nos últimos anos, a literatura vem apresentando a geometria fractal como uma ferramenta poderosa para a análise de imagens, com aplicações em variados campos da ciência. A maior parte destes trabalhos faz uso direto da dimensão fractal como um descritor do objeto representado na imagem. Entretanto, em função da complexidade de muitos problemas nesta área, algumas soluções foram propostas para melhorar essa análise, usando não apenas o valor da dimensão fractal, mas um conjunto de medidas que pudessem ser extraídas pela geometria fractal e que descrevessem as texturas com maior riqueza e precisão. Entre essas técnicas, destacam-se a metodologia de multifractais, de dimensão fractal multiescala e, mais recentemente, os descritores fractais. Esta última técnica tem se mostrado eficiente na solução de problemas relacionados à discriminação de imagens de texturas e formas, uma vez que os descritores gerados fornecem uma representação direta do padrão de complexidade (distribuição dos detalhes ao longo das escalas de observação) da imagem. Assim, essa solução permite que se tenha uma descrição rica da imagem estudada pela análise da distribuição espacial e/ou espectral dos pixels e intensidade de cores/tons de cinza, com uma modelagem que pode se aproximar da percepção visual humana para a geração de um método automático e preciso. Ocorre, entretanto, que os trabalhos apresentados até o momento sobre descritores fractais focam em métodos de estimativa de dimensão fractal mais conhecidos como Bouligand-Minkowski e Box-counting. Este projeto visa estudar mais a fundo o conceito, generalizando para outras abordagens de dimensão fractal, bem como explorando diferentes formas de se extraírem os descritores a partir da curva logarítmica associada à dimensão. Os métodos desenvolvidos são aplicados à análise de texturas, em problemas de classificação de bases públicas, cujos resultados podem ser comparados com métodos da literatura, bem como a segmentação de imagens de satélite e à identificação automática de amostras obtidas em estudos de nanotecnologia. Os resultados alcançados demonstram o potencial da metodologia desenvolvida para a solução destes problemas, mostrando tratar-se de uma nova fronteira a ser usada e explorada em análise de imagens e visão computacional como um todo. / This project describes the development, study and application of fractal descriptors to texture analysis. Recently, the literature has shown fractal geometry as a powerful tool for image analysis, with applications to several areas of science. Most of these works use fractal dimension as a descriptor of the object depicted in the image. However, due to the complexity of many problems in this context, some solutions have been proposed to improve this analysis. These proposed methods use not only the value of fractal dimension, but a set of measures which could be extracted by fractal geometry to describe the textures with greater richness and accuracy. Among such techniques, we emphasize the multifractal methodology, multiscale fractal dimension and, more recently, fractal descriptors. This latter technique has demonstrated to be efficient in solving problems related to the discrimination of texture and shape images. This is possible as the extracted descriptors provide a direct representation of the complexity (the details distribution along the scales of observation) in the image. Thus, this solution allows for a rich description of the image studied by analyzing the spatial/spectral distribution of pixels and intensity of colors/gray-levels, with a model which can approximate the human visual perception, generating an automatic and precise method. However, the works about fractal descriptors presented in the literature focus on classical methods to estimate fractal dimension, such as Bouligand-Minkowski and Box-counting. This project aims at studying more deeply the concept, generalizing to other approaches in fractal dimension, as well as exploring different ways of extracting the key features from the logarithmic curve associated with the dimension. The developed methods are applied to texture analysis, in classification problems over public databases, whose results can be compared with literature methods, as well as to the segmentation of satellite images and automatically identifying samples obtained from studies on nanotechnology. The results demonstrate the potential of the methodology developed to solve such problems, showing that this is a new frontier to be explored and used in image analysis and computer vision at all.

Orientadores: Konradin Metze, Irene Gyongyver Heidemarie Lorand-Metze / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Ciências Médicas / Made available in DSpace on 2018-08-16T09:21:04Z (GMT). No. of bitstreams: 1 Ferro_DanielaPeixoto_M.pdf: 1415087 bytes, checksum: 6c36767b70dc498a2f9f0496b243498a (MD5) Previous issue date: 2010 / Resumo: Características da textura nuclear, realizadas por análise de imagem computadorizada, tem proporcionado informação prognóstica importante em várias neoplasias. Recentemente, a dimensão fractal (DF) da cromatina tem se mostrado um fator independente de prognóstico na leucemia linfóide aguda, no melanona maligno, em carcinomas epidermóides orais e linfomas.Neste estudo nós investigamos a influência da DF da cromatina na sobrevida de pacientes com mieloma múltiplo. Foram estudados 67 pacientes da nossa instituição tratados de acordo com o Grupo de Estudo Brasileiro de mieloma múltiplo. O diagnóstico foi feito pelos critérios do "International Myeloma Working Group". Foi realizado citogenética, eletroforese de proteínas, urina I, com a pesquisa de proteína monoclonal, avaliação da função renal e cálcio sérico. Para o estadiamento, utilizamos o índice prognóstico internacional (ISS). Para cada paciente, foram analisados pelo menos 40 núcleos de esfregaços de medula óssea corados com May-Grünwald-Giemsa. A DF foi determinada com imagens transformadas em escala de cinza pelo método Minkowski-Bouligand estendido para três dimensões. O "goodness-of-fit" da DF foi estimado pelos valores de R² em gráficos log-log. A influência dos parâmetros estudados de sobrevida dos pacientes foi analisada pelos métodos Kaplan-Meier e pela regressão de Cox. A idade média dos pacientes foi de 56 anos. Segundo o ISS, 14% dos pacientes eram do estádio I, 39% eram de estádio II e 47% eram de estádio III. A análise citogenética revelou dois pacientes com alterações do cromossomo 13, dois com translocações envolvendo o cromossomo 14 (em um caso, juntamente com -17) e um paciente com hipodiploidia. Fatores de risco adicional foram encontrados em 62% dos pacientes. Na análise univariada, tanto a DF, quanto o goodness-of-fit foram fatores prognósticos, este último após estratificação pelo ISS. Alta dimensão fractal e baixo "goodness-of-fit" indicaram um pior prognóstico. Na regressão multivariada de Cox, DF, R², ISS e aberrações cromossômicas entraram no modelo final, que mostrou-se estável em um estudo reamostragem bootstrap. Em resumo, as características fractais da cromatina em citologia de rotina revelaram informações relevantes no prognóstico dos doentes com mieloma múltiplo. / Abstract: Nuclear texture features, analyzed by computerized image analysis, has provided important prognostic information in several neoplasias. Recently, the fractal dimension (FD) of the chromatin structure has shown to be an independent prognostic factor in lymphoblastic leukemia acute,in malignant melanoma, in oral squamous cell carcinomas and linfomas. In this study we investigated the influence of the FD of chromatin on survival of patients with multiple myeloma. We studied 67 patients from our Institution treated in the Brazilian Multiple Myeloma Study Group. The diagnosis was confirmed by the criteria of International Myeloma Working Group. Was performed cytogenetic protein electrophoresis, urine I, with the research of protein monoclonal, assessment of renal function and serum calcium.The international Prognostic Index (ISS) was used for staging. For every patient, images of at least 40 nuclei from May-Grünwald-Giemsa stained bone marrow smears were analyzed. FD was determined in gray-scale transformed images by the Minkowski-Bouligand method extended to three dimensions. Goodness-of-fit of FD was estimated by the R2 values in the log-log plots. The influence of parameters studied patients survival was analyzed by Kaplan- Meier and Cox regression. Median age of the patients was 56 years. According to ISS, 14% of the patients were stage I, 39% were stage II and 47% were stage III. Cytogenetic analysis revealed two patients with alterations of chromosome 13, with two translocations involving chromosome 14 (in one case with -17) and one patient with hypodiploid. Additional risk factors were found in 62% of patients. In the univariate analysis FD as well as its goodness-of-fit were prognostic factors, the latter after stratifying for the ISS stage. Higher FD dimension or lower goodness-of-fit indicated a poor prognosis. In the multivariate Cox-regression, FD, R2, ISS stage and chromosomal aberrations entered the final model, which showed to be stable in a bootstrap resampling study. In short, fractal characteristics of the chromatin in routine cytology reveal relevant prognostic information in patients with multiple myeloma. / Mestrado / Biologia Estrutural, Celular, Molecular e do Desenvolvimento / Mestre em Fisiopatologia Médica

A displasia fibrosa é uma alteração de desenvolvimento caracterizada pela substituição do osso normal por tecido conjuntivo denso e trabéculas ósseas imaturas, geralmente encontrada em adolescentes e adultos jovens. Uma alteração genética que envolve a proteína Gs-alfa parece ser a base do processo. A exata incidência e prevalência são difíceis de estabelecer, mas as lesões representam cerca de 5% a 7% dos tumores ósseos benignos. Nos ossos craniofaciais tem predileção pela maxila, podendo causar deformidade grave e assimetria, afetando igualmente ambos os sexos. Radiograficamente, pode apresentar diferentes padrões de imagem dependendo do grau de mineralização e maturação da lesão. .A avaliação da displasia fibrosa nas radiografias da região craniofacial pode ser difícil por causa das aparências variáveis e das estruturas que se sobrepõem, de modo que a tomografia computadorizada é um recurso relevante para o seu correto diagnóstico e planejamento de tratamento. O objetivo deste estudo foi caracterizar a displasia fibrosa através da análise da lacunaridade, um método multiescala para descrever padrões de dispersão espacial. Foram avaliados 10 pacientes (6 homens e 4 mulheres) comprometendo a maxila em sua grande maioria. Para a análise da lacunaridade, empregou-se cortes tomográficos axiais e coronais e, posteriormente, selecionou-se as regiões de interesse das áreas displásicas e do osso normal contralateral por meio do software MATLAB®. Após testes e análises estatísticas, concluiu-se que os cortes coronais, com ampliação de 3x do seu tamanho original, mostraram superioridade em relação aos axiais e, que a lacunaridade foi menor nas áreas da região displásica em relação ao osso normal, ou seja, a primeira apresentou uma maior homogeneidade de textura que a segunda. Mediante isso, pela técnica da validação cruzada leave-one-out é possível separar os grupos com uma alta acurácia (94,75%) concluindo-se que a lacunaridade é um método de análise de imagens contributivo na caracterização da displasia fibrosa. / Fibrous dysplasia is an alteration of development characterized by replacing normal bone for dense connective tissue and immature trabecular bones, typically found in teenagers and young adults. Genetic modification which involves alpha-Gs protein appears to be the basis of the process. The exact incidence and prevalence are difficult to be established, but injuries represent about 5% to 7% of benign bone tumors. On the craniofacial bones, the tumors have a predilection for the maxilla and often can cause severe deformity and asymmetry affecting both sexes equally. Radiographically, it may have different patterns depending on the image degree of mineralization and maturation of the lesion. The evaluation of radiographs of fibrous dysplasia in the craniofacial region can be difficult because of the different appearances and structures that overlaps, however, CT is an important resource for proper diagnosis and treatment planning. The aim of this study was to characterize the fibrous dysplasia by analyzing the lacunarity which is a multiscale method to describe patterns of spatial dispersion. We evaluated 10 patients (6 males and 4 females) and the maxillary was the most affected area. To the lacunarity analysis, we used an axial and coronal view and then were selected the regions of interest in the areas of dysplastic and contralateral normal bone by means of MATLAB® software. After tests and statistical analysis can be conclued that the coronal magnification 3x its original size showed superiority compared to thrust, and that the lacunarity was lower in the areas of dysplastic region in relation to normal bone, namely the first presented a more uniform texture than the second. Through this, the technique of cross-validation \"leave-one-out\" is possible to separate the groups with a high accuracy (94.75%) concluding that the lacunarity is a method of image analysis to characterize the contributory fibrous dysplasia.

A análise de texturas tem atraído um crescente interesse em visão computacional devido a sua importância na caracterização de imagens. Basicamente, as pesquisas em texturas podem ser divididas em duas categorias: texturas estáticas e texturas dinâmicas. As texturas estáticas são caracterizadas por variações de intensidades que formam um determinado padrão repetido espacialmente na imagem. Por outro lado, as texturas dinâmicas são padrões de texturas presentes em uma sequência de imagens. Embora muitas pesquisas tenham sido realizadas, essa área ainda se encontra aberta a estudos, principalmente em texturas dinâmicas por se tratar de um assunto recente e pouco explorado. Este trabalho tem como objetivo o desenvolvimento de pesquisas que abrangem ambos os tipos de texturas nos âmbitos teórico e prático. Em texturas estáticas, foram propostos dois métodos: (i) baseado em caminhadas determinísticas parcialmente auto-repulsivas e dimensão fractal - (ii) baseado em atividade em redes direcionadas. Em texturas dinâmicas, as caminhadas determinísticas parcialmente auto-repulsivas foram estendidas para sequências de imagens e obtiveram resultados interessantes em reconhecimento e segmentação. Os métodos propostos foram aplicados em problemas da biologia e nanotecnologia, apresentando resultados interessantes para o desenvolvimento de ambas as áreas. / Texture analysis has attracted an increasing interest in computer vision due to its importance in describing images. Basically, research on textures can be divided into two categories: static and dynamic textures. Static textures are characterized by intensity variations which form a pattern repeated in the image spatially. On the other hand, dynamic textures are patterns of textures present in a sequence of images. Although many studies have been carried out, this area is still open to study, especially in dynamic textures since it is a recent and little-explored subject. This study aims to develop research covering both types of textures in theoretical and practical fields. In static textures, two methods were proposed: (i) based on deterministic partially self-avoiding walks and fractal dimension - (ii) based on activity in directed networks. In dynamic textures, deterministic partially self-avoiding walks were extended to sequences of images and obtained interesting results in recognition and segmentation. The proposed methods were applied to problems of biology and nanotechnology, presenting interesting results in the development of both areas.

Orientadores: Evandro Pinto da Luz de Oliveira, Gabriela Castellano / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Ciências Médicas / Made available in DSpace on 2018-08-23T10:45:24Z (GMT). No. of bitstreams: 1 Baldissin_MauricioMartins_M.pdf: 5835466 bytes, checksum: 04b04142d756e7073a038bb534379770 (MD5) Previous issue date: 2013 / Resumo: As epilepsias refratárias compreendem síndromes para as quais as terapias que empregam duas ou mais drogas antiepilépticas (DAEs), isoladamente ou em associação, não resultam no controle da frequência das crises. Portadores podem apresentar displasias corticais focais (DCFs) ou difusas e/ou alterações atróficas hipocampaise,em alguns casos, não são detectáveis por uma simples análise visual nas imagens de ressonância magnética (RM). Nesse contexto, o objetivo deste estudo foi avaliar a textura de imagens de RM em regiões de interesse (ROIs) localizadas nos hipocampos, córtex de associação límbico e córtex pré-frontal de 20 pacientes com epilepsia refratária e compará-las às mesmas áreas de um grupo de 20 sujeitos sadios. As imagens de RM desses pacientes não apresentavam alterações visualmente detectáveis. A abordagem utilizada para a estimativa de parâmetros de textura foi a matriz de coocorrência de níveis de cinza (MCO). Dos 11 parâmetros de textura calculados, sete indicaram a existência de diferenças estatisticamente significantes entre pacientes e controles. Tais achados sugerem que a técnica de análise de textura pode contribuir para os estudos das epilepsias refratárias, e possui potencial para servir de auxílio no diagnóstico destas síndromes / Abstract: Refractory epilepsies are syndromes for which the therapies that use two or more antiepileptic drugs (AEDs), in a single way or together, do not result in the control of the frequency of the crises. The patients can present diffuse or focal cortical dysplasia (FCDs) and/or hippocampus atrophic changes that, in some cases, are not detectable by a simple visual analysis of the magnetic resonace (MR) images. In this context, the objective of this study was to assess the MR images texture in regions of interest (ROIs) placed in the hippocampi, limbic association cortex and prefrontal cortex of 20 patients with refractory epilepsy, and compare them with the same areas of a group of 20 healthy individuals. The MR images of these patients did not present visually detectable changes. The approach used for estimating the texture parameters was the gray level coocurrence matrix. Out of the 11 texture parameters calculated, seven indicated the existence of statistically significant differences among patients and controls. These findings suggest that the technique of texture analysis can contribute for the study of refractory epilepsies, and has potential to serve as an aid in the diagnosis of these syndromes / Mestrado / Neurologia / Mestre em Ciências Médicas

Abstract Texture plays an important role in numerous computer vision applications. Many methods for describing and analyzing of textured surfaces have been proposed. Variations in the appearance of texture caused by changing illumination and imaging conditions, for example, set high requirements on different analysis methods. In addition, real-world applications tend to produce a great deal of complex texture data to be processed that should be handled effectively in order to be exploited. A local binary pattern (LBP) operator offers an efficient way of analyzing textures. It has a simple theory and combines properties of structural and statistical texture analysis methods. LBP is invariant against monotonic gray-scale variations and has also extensions to rotation invariant texture analysis. Analysis of real-world texture data is typically very laborious and time consuming. Often there is no ground truth or other prior knowledge of the data available, and important properties of the textures must be learned from the images. This is a very challenging task in texture analysis. In this thesis, methods for learning and recognizing texture categories using local binary pattern features are proposed. Unsupervised clustering and dimensionality reduction methods combined to visualization provide useful tools for analyzing texture data. Uncovering the data structures is done in an unsupervised fashion, based only on texture features, and no prior knowledge of the data, for example texture classes, is required. In this thesis, non-linear dimensionality reduction, data clustering and visualization are used for building a labeled training set for a classifier, and for studying the performance of the features. The thesis also proposes a multi-class approach to learning and labeling part based texture appearance models to be used in scene texture recognition using only little human interaction. Also a semiautomatic approach to learning texture appearance models for view based texture classification is proposed. The goal of texture characterization is often to classify textures into different categories. In this thesis, two texture classification systems suitable for different applications are proposed. First, a discriminative classifier that combines local and contextual texture information of the image in scene recognition is proposed. Secondly, a real-time capable texture classifier with a self-intuitive user interface to be used in industrial texture classification is proposed. Two challenging real-world texture analysis applications are used to study the performance and usefulness of the proposed methods. The first one is visual paper analysis which aims to characterize paper quality based on texture properties. The second application is outdoor scene image analysis where texture information is used to recognize different regions in the scenes.

Land cover mapping is an important part of resource management, planning, and economic predictions. Improvements in remote sensing, machine learning, image processing, and object based image analysis (OBIA) has made the process of identifying land cover types increasingly faster and reliable but these advances are unable to utilize the amount of information encompassed within ultra-high (sub-meter) resolution imagery. Previously, users have typically reduced the resolution of imagery in an attempt to more closely represent the interpretation or object scale in an image and rid the image of any extraneous information within the image that may cause the OBIA process to identify too small of objects when performing semi-automated delineation of objects based on an images’ properties (Mas et al., 2015; Eiesank et al., 2014; Hu et al., 2010). There have been few known attempts to try and maximize this detailed information in high resolution imagery using advanced textural components. In this study we try to circumnavigate the inherent problems associated with high resolution imagery by combining well researched data transformations that aid the OBIA process with a seldom used texture transformation in Geographic Object Based Image Analyses (GEOBIA) known as the Gabor Transform and the hierarchal organization of landscapes. We will observe the difference made in segmentation and classification accuracy of a random forest classifier when we fuse a Gabor transformed image to a Normalized Difference Vegetation Index (NDVI), high resolution multi-spectral imagery (RGB and NIR) and Light Detection and Ranging (LiDAR) derived canopy height model (CHM) within a riparian area in Southeast Iowa. Additionally, we will observe the effects on classification accuracy when adding multi-scale land cover data to objects. Both, the addition of hierarchical information and Gabor textural information, could aid the GEOBIA process in delineating and classifying the same objects that human experts would delineate within this riparian landscape.

The thesis deals with the detection of the nerve fiber layer disruptions in retina scans. The introduction presents an overview of the human eye fysiology and analyses the input image data. The thesis continues with an investigation of two texture analysis methods. While the method of adapted filters does not produce very good results, the method of brightness assessment is shown to work satisfactorily. The final part of the thesis describes the implemented tool for the detection of the nerve fiber layer disruptions.

Dénommée multiparamétrique par adjonction de séquences fonctionnelle aux conventionnelles, L’IRM de prostate a montré ses performances pour la détection du cancer de prostate par un score radiologique visuel, subjectif. D’autres applications sont en cours d’investigations comme la stratification, le planning thérapeutique ou encore le suivi oncologique.La première partie s’attache à décrire, élaborer et appliquer une méthodologie de recalage non rigide en 3D entre l’histologie du spécimen de prostatectomie totale et les différentes séquences de l’IRMmp. Après avoir capturé une déformation et un changement de volume de la prostate entre les états in vivo et ex vivo par IRM, la méthode de recalage multimodalité appliquée à une population de prostatectomie totale précédée d’une IRM démontre une sous-estimation du volume de cancer par l’IRM, sujette à une stratification. Les implications se trouvent dans la détection, la stratification et le planning thérapeutique. La deuxième partie propose une analyse de texture des différentes séquences et cartographies quantitatives en diffusion et perfusion pour la détection et la stratification du cancer. Cette approche multiparamétrique de « Score d’Entropie » est testée dans une population pilote au moment des biopsies et présente des performances diagnostique pour sélectionner les lésions à biopsier. Ce score d’entropie participe de la stratification du cancer en corrélant positivement avec le score de Gleason et la longueur de cancer biopsique.La troisième partie explore le rôle de l’IRM dans le suivi d’une thérapie émergente, dite focale, du cancer. Il s’agit d’un travail de recalage non-rigide longitudinal sur une cohorte de patients traités par thérapie focale en vue de compenser les déformation focalement induites. Il apparaît que ce type de recalage peut permettre un suivi objectif des résultats d’ablation et potentiellement élaborer une cible biopsique et radiologique dans le suivi oncologique. / Conventional prostate MRI, enhanced by diffusion and perfusion sequences, and then named multiparametric, showed high performances for detection of prostate cancer using visual scoring. Indications in stratification, prognosis, treatment planning and follow up are currently under investigations.First part of this work attached itself to describe, elaborate and use a non-rigid image fusion method in 3D between gold standard histology of radical prostatectomy and MRI. Investigations captured the significant differences in shape and volume of in vivo and ex vivo prostate using MRI. The developed multimodality fusion method was applied to a cohort of patients who underwent MRI prior surgery. Results showed a stratified underestimation of cancer volume by MRI. Clinical output resides in detection, stratification and surgical planning.The second part proposed some texture analysis of sequences and quantitative maps. As a multiparametric approach, the Entropy Score is applied in a pilot cohort at time of biopsy and showed some potential usefulness to select MRI targets without compromising detection of significant cancer. By positively correlating with the Gleason Score and the maximal core length of cancer, Entropy Score participates of stratification of cancer.The third part explored application of image registration in the longitudinal follow up of an emergent therapy, said focal (FT). As a conservative approach, FT induces very local deformation of the gland which appears to be appropriately modelled by non-rigid registration, then opening possibilities to guide further control biopsy and radiologic assessment.