The Ising model is an important model in statistical physics, with over 10,000 papers published on the topic. This model assumes binary variables and only local pairwise interactions between neighbouring nodes. Inference for the general Ising model is NP-hard; this includes tasks such as calculating the partition function, finding a lowest-energy (ground) state and computing marginal probabilities. Past approaches have proceeded by working with classes of tractable Ising models, such as Ising models defined on a planar graph. For such models, the partition function and ground state can be computed exactly in polynomial time by establishing a correspondence with perfect matchings in a related graph. In this thesis we continue this line of research. In particular we simplify previous inference algorithms for the planar Ising model. The key to our construction is the complementary correspondence between graph cuts of the model graph and perfect matchings of its expanded dual. We show that our exact algorithms are effective and efficient on a number of real-world machine learning problems. We also investigate heuristic methods for approximating ground states of non-planar Ising models. We show that in this setting our approximative algorithms are superior than current state-of-the-art methods.

Dans ce travail nous nous intéressons à la segmentation d’image 3D. Le but est de définir un cadre permettant, étant donnée une problématique de segmentation, de développer rapidement un algorithme apportant une solution à cette problématique. Afin de ne pas être restreint à un sous ensemble des types de problématique de segmentation, ce cadre doit permettre de mettre en oeuvre efficacement les différentes méthodes et les différents critères de segmentation existants, dans le but de les combiner pour définir les nouveaux algorithmes. Ce cadre doit reposer sur un modèle de structuration d’image qui représente la topologie et la géométrie d’une partition et permet d’en extraire efficacement les informations requises. Dans ce document, les différentes méthodes de segmentation existantes sont présentées afin de définir un ensemble d’opération nécessaire à leur implémentation. Une présentation des modèles existants est faite pour en déterminer avantages et inconvénients, puis le nouveau modèle est ensuite défini. Sa mise en oeuvre complète est détaillée ainsi qu’une analyse de sa complexité en temps et en mémoire pour l’ensemble des opérations précédemment définies. Des exemples d’utilisation du modèle sur des cas concrets sont ensuite décrits, ainsi que les possibilités d’extension du modèle et d’implémentation sur architecture parallèle. / In this work we focus on 3D image segmentation. The aim consists in defining a framework which, given a segmentation problem, allows to design efficiently an algorithm solving this problem. Since this framework has to be unspecific according to the kind of segmentation problem, it has to allow an efficient implementation of most segmentation techniques and criteria, in order to combine them to define new algorithms. This framework has to rely on a structuring model both representing the topology and the geometry of the partition of an image, in order to efficiently extract required information. In this document, different segmentation techniques are presented in order to define a set of primitives required for their implementation. Existing models are presented with their advantages and drawbacks, then the new structuring model is defined. Its whole implementation including details of its memory consumption and time complexity for each primitives of the previously defined set of requirements is given. Some examples of use with real image analysis problems are described, with also possible extensions of the model and its implementation on parallel architecture.

abstract: Spotlight mode synthetic aperture radar (SAR) imaging involves a tomo- graphic reconstruction from projections, necessitating acquisition of large amounts of data in order to form a moderately sized image. Since typical SAR sensors are hosted on mobile platforms, it is common to have limitations on SAR data acquisi- tion, storage and communication that can lead to data corruption and a resulting degradation of image quality. It is convenient to consider corrupted samples as missing, creating a sparsely sampled aperture. A sparse aperture would also result from compressive sensing, which is a very attractive concept for data intensive sen- sors such as SAR. Recent developments in sparse decomposition algorithms can be applied to the problem of SAR image formation from a sparsely sampled aperture. Two modified sparse decomposition algorithms are developed, based on well known existing algorithms, modified to be practical in application on modest computa- tional resources. The two algorithms are demonstrated on real-world SAR images. Algorithm performance with respect to super-resolution, noise, coherent speckle and target/clutter decomposition is explored. These algorithms yield more accu- rate image reconstruction from sparsely sampled apertures than classical spectral estimators. At the current state of development, sparse image reconstruction using these two algorithms require about two orders of magnitude greater processing time than classical SAR image formation. / Dissertation/Thesis / M.S. Electrical Engineering 2011

No description available.

[pt] A segmentação de imagens com objetivo de determinar a forma de objetos é ainda um problema difícil. A separação de regiões que correspondem a objetos contidos na imagem geralmente leva em consideração propriedades de similaridade, proximidade e descontinuidade. A imagem a ser segmentada pode ser de diversas naturezas, como fotografias, imagens médicas e sísmicas. Podemos encontrar na literatura muitos métodos de segmentação propostos como possíveis soluções para diferentes problemas. Recentemente a técnica de superpixel tem sido utilizada como um passo inicial que reduz o tamanho da entrada do problema. Este trabalho propõe uma metodologia de segmentação de imagens fotográficas e de ultrassom que se baseia em variantes de superpixels. A metodologia proposta se adapta a natureza da imagem e a complexidade do problema utilizando diferentes medidas de similaridade e distância. O trabalho apresenta também resultados que buscam esclarecer o procedimento proposto e a escolha de seus parâmetros. / [en] Image segmentation for object modeling is a complex task that is still not well solved. The separation of the regions corresponding to each object in an image is based on proximity, similarity, and discontinuity of its boundaries. The image to be segmented can be of various natures, including photographs, medical and seismic images. We can find in literature many proposed segmentation methods used as solutions to different problems. Recently the superpixel technique has been used as an initial step that reduces the size of the problem input. This work proposes a methodology of segmentation of photographs and ultrasound images based on variants of superpixels. The proposed methodology adapts to the image s nature and to the problem s complexity using different measures of similarity and distance. This work also presents results that seek to clarify the proposed procedure and the choice of its parameters.

A segmentação de imagens consiste em urna tarefa de fundamental importância para diferentes aplicações em visão computacional, tais como por exemplo, o reconhecimento e o rastreamento de objetos, a segmentação de tomores/lesões em aplicações médicas, podendo também servir de auxílio em sistemas de reconhecimento facial. Embora exista uma extensa literatora abordando o problema de segmentação de imagens, tal tópico ainda continua em aberto para pesquisa. Particularmente, a tarefa de segmentar imagens coloridas é desafiadora devido as diversas inomogeneidades de cor, texturas e formas presentes nas feições descritivas das imagens. Este trabalho apresenta um novo método de clustering para abordar o problema da segmentação de imagens coloridas. Nós desenvolvemos uma abordagem Bayesiana para procura de máximos de densidade em urna distribuição discreta de dados, e representamos os dados de forma hierárquica originando clusters adaptativos a cada nível da hierarquia. Nós aplicamos o método de clustering proposto no problema de segmentação de imagens coloridas, aproveitando sua estrutura hierárquica, baseada em propriedades de árvores direcionadas, para representar hierarquicamente uma imagem colorida. Os experimentos realizados revelaram que o método de clustering proposto, aplicado ao problema de segmentação de imagens coloridas, obteve para a medida de performance Probabilistic Rand lndex (PRI) o valor de 0.8148 e para a medida Global Consistency Error (GCE) o valor 0.1701, superando um total de vinte e um métodos previamente propostos na literatura para o banco de dados BSD300. Comparações visuais confirmaram a competitividade da nossa abordagem em relação aos demais métodos testados. Estes resultados enfatizam a potencialidade do nosso método de clustering para abordar outras aplicações no domínio de Visão Computacional e Reconhecimento de Padrões. / Image segmentation is an essential task for several computer vision applications, such as object recognition, tracking and image retrieval. Although extensively studied in the literature, the problem of image segmentation remains an open topic of research. Particularly, the task of segmenting color images is challenging due to the inhomogeneities in the color regions encountered in natural scenes, often caused by the shapes of surfaces and their interactions with the illumination sources (e.g. causing shading and highlights) This work presents a novel non-supervised classification method. We develop a Bayesian framework for seeking modes on the underlying discrete distribution of data and we represent data hierarchically originating adaptive clusters at each levei of hierarchy. We apply the prnposal clustering technique for tackling the problem of color irnage segmentation, taking advantage of its hierarchical structure based on hierarchy properties of directed trees for representing fine to coarse leveis of details in an image. The experiments herein conducted revealed that the proposed clustering method applied to the color image segmentation problem, achieved for the Probabilistic Rand Index (PRI) performance measure the value of 0.8148 and for the Global Consistency Error (GCE) the value of 0.1701, outperforming twenty-three methods previously proposed in the literature for the BSD300 dataset. Visual comparison confirmed the competitiveness of our approach towards state-of-art methods publicly available in the literature. These results emphasize the great potential of our proposed clustering technique for tackling other applications in computer vision and pattem recognition.

Changes in retinal structure are related to different eye diseases. Various retinal imaging techniques, such as fundus imaging and optical coherence tomography (OCT) imaging modalities, have been developed for non-intrusive ophthalmology diagnoses according to the vasculature changes. However, it is time consuming or even impossible for ophthalmologists to manually label all the retinal structures from fundus images and OCT images. Therefore, computer aided diagnosis system for retinal imaging plays an important role in the assessment of ophthalmologic diseases and cardiovascular disorders. The aim of this PhD thesis is to develop segmentation methods to extract clinically useful information from these retinal images, which are acquired from different imaging modalities. In other words, we built the segmentation methods to extract important structures from both 2D fundus images and 3D OCT images. In the first part of my PhD project, two novel level set based methods were proposed for detecting the blood vessels and optic discs from fundus images. The first one integrates Chan-Vese's energy minimizing active contour method with the edge constraint term and Gaussian Mixture Model based term for blood vessels segmentation, while the second method combines the edge constraint term, the distance regularisation term and the shape-prior term for locating the optic disc. Both methods include the pre-processing stage, used for removing noise and enhancing the contrast between the object and the background. Three automated layer segmentation methods were built for segmenting intra-retinal layers from 3D OCT macular and optic nerve head images in the second part of my PhD project. The first two methods combine different methods according to the data characteristics. First, eight boundaries of the intra-retinal layers were detected from the 3D OCT macular images and the thickness maps of the seven layers were produced. Second, four boundaries of the intra-retinal layers were located from 3D optic nerve head images and the thickness maps of the Retinal Nerve Fiber Layer (RNFL) were plotted. Finally, the choroidal layer segmentation method based on the Level Set framework was designed, which embedded with the distance regularisation term, edge constraint term and Markov Random Field modelled region term. The thickness map of the choroidal layer was calculated and shown.

Master of Science / Department of Biological & Agricultural Engineering / Ajay Sharda / Precise water application conserves resources, reduces costs, and optimizes plant performance and quality. Existing irrigation scheduling utilizes single, localized measurements that do not account for spatial crop water need; but, quick, single-point sensors are impractical for measuring discrete variations across large coverage areas. Thermography is an alternate approach for measuring spatial temperatures to quantify crop health. However, agricultural studies using thermography are limited due to previous camera expense, unfamiliar use and calibration, software for image acquisition and high-throughput processing specifically designed for thermal imagery mapping and monitoring spatial crop water need. Recent advancements in thermal detectors and sensing platforms have allowed uncooled thermal infrared (TIR) cameras to become suited for crop sensing. Therefore, a small, lightweight thermal infrared imaging system (TIRIS) was developed capable of radiometric temperature measurements. One-time (OT) and real-time (RT) radiometric calibrations methods were developed and validated for repeatable, temperature measurements while compensating for strict environmental conditions within a climate chamber. The Tamarisk® 320 and 640 analog output yielded a measurement accuracy of ±0.82°C or 0.62ºC with OT and RT radiometric calibration, respectively. The Tamarisk® 320 digital output yielded a measurement accuracy of ±0.43 or 0.29ºC with OT and RT radiometric calibration, respectively. Similarly, the FLIR® Tau 2 analog output yielded a measurement accuracy of ±0.87 or 0.63ºC with OT and RT radiometric calibration, respectively. A TIRIS was then built for high-throughput image capture, correction, and processing and RT environmental compensation for monitoring crop water stress within a greenhouse and temperature mapping aboard a small unmanned aerial systems (sUAS). The greenhouse TIRIS was evaluated by extracting plant temperatures for monitoring full-season crop water stress index (CWSI) measurements. Canopy temperatures demonstrated that CWSI explained 82% of the soil moisture variation. Similarly, validation aboard a sUAS provided radiometric thermal maps with a ±1.38°C (α=0.05) measurement accuracy. Due to the TIR cameras’ performance aboard sUAS and greenhouse platforms, a TIRIS provides unparalleled spatial coverage and measurement accuracy capable of monitoring subtle crop stress indicators. Further studies need to be conducted to produce spatial crop water stress maps at scales necessary for variable rate irrigation systems.

abstract: Accurate quantitative information of tumor/lesion volume plays a critical role in diagnosis and treatment assessment. The current clinical practice emphasizes on efficiency, but sacrifices accuracy (bias and precision). In the other hand, many computational algorithms focus on improving the accuracy, but are often time consuming and cumbersome to use. Not to mention that most of them lack validation studies on real clinical data. All of these hinder the translation of these advanced methods from benchside to bedside. In this dissertation, I present a user interactive image application to rapidly extract accurate quantitative information of abnormalities (tumor/lesion) from multi-spectral medical images, such as measuring brain tumor volume from MRI. This is enabled by a GPU level set method, an intelligent algorithm to learn image features from user inputs, and a simple and intuitive graphical user interface with 2D/3D visualization. In addition, a comprehensive workflow is presented to validate image quantitative methods for clinical studies. This application has been evaluated and validated in multiple cases, including quantifying healthy brain white matter volume from MRI and brain lesion volume from CT or MRI. The evaluation studies show that this application has been able to achieve comparable results to the state-of-the-art computer algorithms. More importantly, the retrospective validation study on measuring intracerebral hemorrhage volume from CT scans demonstrates that not only the measurement attributes are superior to the current practice method in terms of bias and precision but also it is achieved without a significant delay in acquisition time. In other words, it could be useful to the clinical trials and clinical practice, especially when intervention and prognostication rely upon accurate baseline lesion volume or upon detecting change in serial lesion volumetric measurements. Obviously, this application is useful to biomedical research areas which desire an accurate quantitative information of anatomies from medical images. In addition, the morphological information is retained also. This is useful to researches which require an accurate delineation of anatomic structures, such as surgery simulation and planning. / Dissertation/Thesis / Doctoral Dissertation Biomedical Informatics 2016

La délimitation précise du contour des lésions pigmentées sur des images est une première étape importante pour le diagnostic assisté par ordinateur du mélanome. Cette thèse présente une nouvelle approche de la détection automatique du contour des lésions pigmentaires sur des images couleurs ou multispectrales de la peau. Nous présentons d'abord la notion de minimisation d'énergie par coupes de graphes en terme de Maxima A-Posteriori d'un champ de Markov. Après un rapide état de l'art, nous étudions l'influence des paramètres de l'algorithme sur les contours d'images couleurs. Dans ce cadre, nous proposons une fonction d'énergie basée sur des classifieurs performants (Machines à support de vecteurs et Forêts aléatoires) et sur un vecteur de caractéristiques calculé sur un voisinage local. Pour la segmentation de mélanomes, nous estimons une carte de concentration des chromophores de la peau, indices discriminants du mélanomes, à partir d'images couleurs ou multispectrales, et intégrons ces caractéristiques au vecteur. Enfin, nous détaillons le schéma global de la segmentation automatique de mélanomes, comportant une étape de sélection automatique des "graines" utiles à la coupure de graphes ainsi que la sélection des caractéristiques discriminantes. Cet outil est comparé favorablement aux méthodes classiques à base de coupure de graphes en terme de précision et de robustesse. / Accurate border delineation of pigmented skin lesion (PSL) images is a vital first step in computer-aided diagnosis (CAD) of melanoma. This thesis presents a novel approach of automatic PSL border detection on color and multispectral skin images. We first introduce the concept of energy minimization by graph cuts in terms of maximum a posteriori estimation of a Markov random field (MAP-MRF framework). After a brief state of the art in interactive graph-cut based segmentation methods, we study the influence of parameters of the segmentation algorithm on color images. Under this framework, we propose an energy function based on efficient classifiers (support vector machines and random forests) and a feature vector calculated on a local neighborhood. For the segmentation of melanoma, we estimate the concentration maps of skin chromophores, discriminating indices of melanomas from color and multispectral images, and integrate these features in a vector. Finally, we detail an global framework of automatic segmentation of melanoma, which comprises two main stages: automatic selection of "seeds" useful for graph cuts and the selection of discriminating features. This tool is compared favorably to classic graph-cut based segmentation methods in terms of accuracy and robustness.