Ultrasonography is widely used in medical diagnosis with the advantages of being low cost, non-invasive and capable of real time imaging. When interpreting ultrasonographic images of mammalian ovaries, the structures of interest are follicles, corpora lutea (CL) and stroma. This thesis presents an approach to perform CL texture analysis, including detection and segmentation, based on the classiers trained by genetic programming (GP). The objective of CL detection is to determine whether there is a CL in the ovarian images, while the goal of segmentation is to localize the CL within the image. Genetic programming (GP) oers a solution through the evolution of computer programs by methods inspired by the mechanisms of natural selection. Herein, we use rotationally invariant local binary patterns (LBP) to encode the local texture features. These are used by the programs which are manipulated by GP to obtain highly t CL classiers. Grayscale standardization was performed on all images in our data set based on the reference grayscale in each image. CL classication programs were evolved by genetic programming and tested on ultrasonographic images of ovaries. On the bovine dataset, our CL detection algorithm is reliable and robust. The detection algorithm correctly determined the presence or absence of a CL in 93.3% of 60 test images. The segmentation algorithm achieved a mean ( standard deviation) sensitivity and specicity of 0.87 (0.14) and 0.91 (0.05), respectively, over the 30 CL images. Our CL segmentation algorithm is an improvement over the only previously published algorithm, since our method is fully automatic and does not require the placement of an initial contour. The success of these algorithms demonstrates that similar algorithms designed for analysis of in vivo human ovaries are likely viable.

El análisis de texturas es un área de estudio interesante con suficiente peso específico dentro de los diferentes campos que componen la visión por ordenador. En este trabajo hemos desarrollado métodos específicos para resolver aspectos importantes de dicha área. El primer acercamiento al tema viene de la mano de un problema de segmentación de un tipo de texturas muy concreto como son las imágenes microscópicas de láminas de mármol. Este primer tipo de imágenes se componen de un conjunto de granos cuyas formas y tamaños sirven a los especialistas para identificar, catalogar y determinar el origen de dichas muestras. Identificar y analizar los granos que componen tales imágenes de manera individual necesita de una etapa de segmentación. En esencia, esto implica la localización de las fronteras representadas en este caso por valles que separan zonas planas asociadas a los granos. De los diferentes métodos estudiados para la detección de dichos valles y para el caso concreto de imágenes petrográficas son los basados en técnicas de morfología matemática los que han dado mejores resultados. Además, la segmentación requiere un filtrado previo para el que se han estudiado nuevamente un conjunto de posibilidades entre las que cabe destacar los algoritmos multirresolución basados en wavelets.El segundo problema que hemos atacado en este trabajo es la clasificación de imágenes de textura. En él también hemos utilizado técnicas multirresolución como base para su resolución. A diferencia de otros enfoques de carácter global que encontramos extensamente en la literatura sobre texturas, nos hemos centrado en problemas donde las diferencias visuales entre las clases de dichas texturas son muy pequeñas. Y puesto que no hemos establecido restricciones fuertes en este análisis, las estrategias desarrolladas son aplicables a un extenso espectro de texturas, como pueden ser las baldosas cerámicas, las imágenes microscópicas de pigmentos de efecto, etc.El enfoque que hemos seguido para la clasificación de texturas implica la consecución de una serie de pasos. Hemos centrado nuestra atención en aquellos pasos asociados con las primeras etapas del proceso requeridas para identificar las características importantes que definen la textura, mientras que la clasificación final de las muestras ha sido realizada mediante métodos de clasificación generales. Para abordar estos primeros pasos dentro del análisis hemos desarrollado una estrategia mediante la cual las características de una imagen se ajustan a un modelo que previamente hemos definido, uno de entre varios modelos que están ordenados por complejidad. Estos modelos están asociados a algoritmos específicos y sus parámetros así como a los cálculos que de ellos se derivan. Eligiendo el modelo adecuado, por tanto, evitamos realizar cálculos que no nos aportan información útil para la clasificación.En un tercer enfoque hemos querido llegar a una descripción de textura que nos permita de forma sencilla su clasificación y su síntesis. Para conseguir este objetivo hemos adoptado por un modelo probabilístico. Dicha descripción de la textura nos permitirá la clasificación a través de la comparación directa de modelos, y también podremos, a partir del modelo probabilístico, sintetizar nuevas imágenes.Para finalizar, comentar que en las dos líneas de trabajo que hemos expuesto, la segmentación y la clasificación de texturas, hemos llegado a soluciones prácticas que han sido evaluadas sobre problemas reales con éxito y además las metodologías propuestas permiten una fácil extensión o adaptación a nuevos casos. Como líneas futuras asociadas a estos temas trataremos por un lado de adaptar la segmentación a imágenes que poco o nada tienen que ver con las texturas, en las que se perseguirá la detección de sujetos y objetos dentro de escenas, como apuntamos más adelante en esta misma memoria. Por otro lado, y relacionado con la clasificación, abordaremos un problema todavía sin solución como es el de la ingeniería inversa en pigmentos de efecto, en otras palabras la determinación de los constituyentes en pinturas metalizadas, y en el que utilizaremos los estudios aquí presentados como base para llegar a una posible solución. / An interesting problem in computer vision is the analysis of texture images. In this work, we have developed specific methods to solve important aspects of this problem. The first approach involves segmentation of a specific type of textures, i.e. those of microscopy images of thin marble sections. These images comprise a pattern of grains whose sizes and shapes help specialists to identify the origin and quality of marble samples. To identify and analyze individual grains in these images represents a problem of image segmentation. In essence, this involves identifying boundary lines represented by valleys which separate flat areas corresponding to grains. Of several methods tested, we found those based on mathematical morphology particularly successful for segmentation of petrographical images. This involves a pre-filtering step for which again several approaches have been explored, including multiresolution algorithms based on wavelets. In the second approach we have also used multiresolution analyses to address the problem of classifying texture images. In contrast to more global approaches found in the literature, we have explored situations where visual differences between textures are rather subtle. Since we have tried to impose relatively few restrictions on these analyses, we have developed strategies that are applicable to a wide range of related texture images, such as images of ceramic tiles, microscopic images of effect pigments, etc.The approach we have used for the classification of texture images involves several technical steps. We have focused our attention in the initial low-level analyses required to identify the general features of the image, whereas the final classification of samples has been performed using generic classification methods. To address the early steps of image analysis, we have developed a strategy whereby the general features of the image fit one of several pre-defined models with increasing levels of complexity. These models are associated to specific algorithms, parameters and calculations for the analysis of the image, thus avoiding calculations that do not provide useful information. Finally, in a third approach we want to arrive to a description of textures in such a way that it should be able to classify and synthesize textures. To reach this goal we adopt a probabilistic model of the texture. This description of the texture allows us to compare textures through comparison of probabilistic models, and also use those probabilities to generate new similar images.In conclusion, we have developed strategies of segmentation and classification of textures that provide solutions to practical problems and are potentially applicable with minor modifications to a wide range of situations. Future research will explore (i) the possibility of adapting segmentation to the analysis of images that do not necessarily involve textures, e.g. localization of subjects in scenes, and (ii) classification of effect pigment images to help identify their components.

The purpose of this thesis is to develop a working methodology to color a grey scale image. This thesis is based on approach of using a colored reference image. Coloring grey scale images has no exact solution till date and all available methods are based on approximation. This technique of using a color reference image for approximating color information in grey scale image is among most modern techniques.Method developed here in this paper is better than existing methods of approximation of color information addition in grey scale images in brightness, sharpness, color shade gradients and distribution of colors over objects.Color and grey scale images are analyzed for statistical and textural features. This analysis is done only on basis of luminance value in images. These features are then segmented and segments of color and grey scale images are mapped on basis of distances of segments from origin. Then chromatic values are transferred between these matched segments from color image to grey scale image.Technique proposed in this paper uses better mechanism of mapping clusters and mapping colors between segments, resulting in notable improvement in existing techniques in this category.

Ovarian follicles in women are fluid-filled structures in the ovary that contain oocytes (eggs). A dominant follicle is physiologically selected and ovulates during the menstrual cycle. We examined the echotexture in ultrasonographic images of the follicle wall of dominant ovulatory follicles in women during natural menstrual cycles and dominant anovulatory follicles which developed in women using oral contraceptives (OC). Texture features of follicle wall regions of both ovulatory and anovulatory dominant follicles were evaluated over a period of seven days before ovulation (natural cycles) or peak estradiol concentrations (OC cycles). Differences in echotexture between the two classes of follicles were found for two co-occurrence matrix derived texture features and two edge-frequency based texture features. Co-occurrence energy and homogeneity were significantly lower for ovulatory follicles while edge density and edge contrast were higher for ovulatory follicles. In the each feature space, the two classes of follicle were adequately separable.</p><p>This thesis employed several statistical approaches to analyses of texture features, such as plotting method and the Mann-Kendall method. A distinct change of feature trend was detected 3 or 4 days before the day of ovulation for ovulatory follicles in the two co-occurrence matrix derived texture features and two edge-frequency-based texture features. Anovulatory follicles, exhibited the biggest variation of the feature value 3 or 4 days before the day on which dominant follicles developed to maximum size. This discovery is believed to correspond to the ovarian follicles responding to system hormonal changes leading to presumptive ovulation.</p>

The area of the digital image processing is getting more attention in the hope that it will increase the accuracy of any scientific measurements, such as in determining an object velocity, temperature, and size. While human vision is excellent to recognize and differentiate objects, it has been proven to be a poor tool when it comes to measure the object performance. One of many digital image processing applications is texture analysis whose purpose is to evaluate image patterns. The purpose of this dissertation is to investigate the use of texture analysis as a tool to micro fiber dispersion measurement. Micro fiber dispersion can be found in many applications such as in paper and industry powder engineering. Three cases related to micro fiber dispersion were investigated in this study. The first case was the experimental study of the dispersion in open water channel. Sets of synthetic fibers were put into water channel to simulate a process that can be found in papermaking industry. The research investigated the effect of three operating parameters: fluid velocity, fiber consistency, and fiber aspect ratio to fiber dispersion. Using two-factorial experimental design technique, the main and interaction effects of these parameters were evaluated. The study found that increasing fluid velocity, fiber aspect ratio, and consistency decreased the dispersion level. The study also found that the effect of individual parameters is more pronounced than the role of the interactive terms on the fiber flocculation. The second case considered was applying the fiber dispersion analysis to computer-synthesized images consisting of different arrangements of fibers. Four sets of sub-cases were presented. These sub-cases were divided based on the fiber-concentrated location and fiber distribution. The use of computer-synthesized images was found to be very useful to simulate real situation during fiber dispersion. The third case investigated the fiber distribution on a dry paper. Images for different types of paper were taken and evaluated to see the dispersion level of each type of paper. It was found that the current texture analysis was applicable to determine the dispersion level for dry papers. While three cases indicated that the texture analysis can be used to investigate the fiber dispersion, the texture analysis used here is not a perfect and universal method and may not be suitable to analyze other types of dispersions. The human vision will always be essential to determine if the texture analysis is applicable to any other problem.

Application of the nonlinear optical microscopy (NLOM) for investigation of biological samples has, to date, primarily focused upon the qualitative analysis of images. The general consensus is that the nonlinear optical (NLO) techniques provide enough bio- chemical information when compared to, for example, visible light microscopy. Herein, it is presented a detailed study where a set of tools for quantitative extraction of infor- mation from NLO images were developed and tested for the analysis of complex tissue assemblies. Two-photon excited autofluorescence (TPEF), second-harmonic generation (SHG), and coherent anti-Stokes Raman scattering (CARS) were used for the charac- terization of atherosclerotic plaques. Our NLO-based image analysis of animal arteries affected by atherosclerotic plaque accumulation revealed that images of the healthy regions of the artery can be readily distinguished by marked differences in morphology, due to a fluorescent signal generated from the presence of generally intact elastic layer. Regions affected by lesions were dominated by lipid-rich cells and collagen fibers; the elastic layer was disrupted and the presence of fluorescent particles were also detected. Next, the potential of using information extracted from NLO images lead us to the development of a new optical index for plaque burden (OIPB). Through the OIPB, it was possible to investigate and to classify the plaque severity regarding the already established and currently used definition during clinical analyses. Extrapolating to and anticipating future applications, several methods for extracting specific information from images acquired by each NLOM modality were developed and tested. Texture analysis, particle-specific features, fractal analysis and directionality of components within the images were successfully adapted and tailored to better extract relevant information from the NLO images. Even though the methods presented in this thesis were mostly tested in images from arterial plaques, there is strong evidence that all tools presented here are capable of tracking changes that occur in many medical conditions and applications.

About 50% of cancer patients are treated with X-ray radiation therapy; however, with current treatment feedback, the effects and the efficacy of the treatment are generally detected several weeks/months after treatment completion. This makes the adjustment of the treatment based on early response, and identification of non-responding patients, nearly impossible. In this thesis a novel method combining optical coherence tomography and a gamut of image analysis methods is explored as a potential approach to detecting tissue variability. Applying texture analysis to the optical coherence tomography images may allow for the tracking of radiation therapy induced cell microstructural changes in cancer patients and help in the adjustment of treatment based on early response.

About 50% of cancer patients are treated with X-ray radiation therapy; however, with current treatment feedback, the effects and the efficacy of the treatment are generally detected several weeks/months after treatment completion. This makes the adjustment of the treatment based on early response, and identification of non-responding patients, nearly impossible. In this thesis a novel method combining optical coherence tomography and a gamut of image analysis methods is explored as a potential approach to detecting tissue variability. Applying texture analysis to the optical coherence tomography images may allow for the tracking of radiation therapy induced cell microstructural changes in cancer patients and help in the adjustment of treatment based on early response.

Positron Emission Tomography (PET) is an imaging modality that has become increasingly beneficial in Radiotherapy by improving treatment planning (1). PET reveals tumor volumes that are not well visualized on computed tomography CT or MRI, recognizes metastatic disease, and assesses radiotherapy treatment (1). It also reveals areas of the tumor that are more radiosensitive allowing for dose painting - a non-homogenous dose treatment across the tumor (1). However, PET is not without limitations. The quantitative unit of PET images, the Standardized Uptake Value (SUV), is affected by many factors such as reconstruction algorithm, patient weight, and tracer uptake time (2). In fact, PET is so sensitive that a patient imaged twice in a single day on the same machine and same protocol will produce different SUV values. The objective of this research was to increase the capabilities of PET by exploring other quantitative PET/CT measures for Radiotherapy treatment applications. The technique of quantitative image feature analysis, nowadays known as radiomics, was applied to PET and CT images. Image features were then extracted from PET/CT images and how the features differed between conventional and respiratory-gated PET/CT images in lung cancer was analyzed. The influence of noise on image features was analyzed by applying uncorrelated, Gaussian noise to PET/CT images and measuring how significantly noise affected features. Quantitative PET/CT measures outside of image feature analysis were also investigated. The correlation of esophageal metabolic tumor volumes (tumor volume demonstrating high metabolic uptake) and endoscopically implanted fiducial markers was studied. It was found that certain image features differed greatly between conventional and respiratory-gated PET/CT. The differences were mainly due to the effect of respiratory motion including affine motion, rotational motion and tumor deformation. Also, certain feature groups were more affected by noise than others. For instance, contour-dependent shape features exhibited the least change with noise. Comparatively, GLSZM features exhibited the greatest change with added noise. Discordance was discovered between the inferior and superior tumor fiducial markers and metabolic tumor volume (MTV). This demonstrated a need for both fiducial markers and MTV to provide a comprehensive view of a tumor. These studies called attention to the differences in features caused by factors such as motion, acquisition parameters, and noise, etc. Investigators should be aware of these effects. PET/CT radiomic features are indeed highly affected by noise and motion. For accurate clinical use, these effects must be account by investigators and future clinical users. Further investigation is warranted towards the standardization of PET/CT radiomic feature acquisition and clinical application.

Abstract Image appearance descriptors are needed for different computer vision applications dealing with, for example, detection, recognition and classification of objects, textures, humans, etc. Typically, such descriptors should be discriminative to allow for making the distinction between different classes, yet still robust to intra-class variations due to imaging conditions, natural changes in appearance, noise, and other factors. The purpose of this thesis is the development and analysis of photometric descriptors for the appearance of real life images. The two application areas included in this thesis are face recognition and texture classification. To facilitate the development and analysis of descriptors, a general framework for image description using statistics of quantized filter bank responses modeling their joint distribution is introduced. Several texture and other image appearance descriptors, including the local binary pattern operator, can be presented using this model. This framework, within which the thesis is presented, enables experimental evaluation of the significance of each of the components of this three-part chain forming a descriptor from an input image. The main contribution of this thesis is a face representation method using distributions of local binary patterns computed in local rectangular regions. An important factor of this contribution is to view feature extraction from a face image as a texture description problem. This representation is further developed into a more precise model by estimating local distributions based on kernel density estimation. Furthermore, a face recognition method tolerant to image blur using local phase quantization is presented. The thesis presents three new approaches and extensions to texture analysis using quantized filter bank responses. The first two aim at increasing the robustness of the quantization process. The soft local binary pattern operator accomplishes this by making a soft quantization to several labels, whereas Bayesian local binary patterns make use of a prior distribution of labelings, and aim for the one maximizing the a posteriori probability. Third, a novel method for computing rotation invariant statistics from histograms of local binary pattern labels using the discrete Fourier transform is introduced. All the presented methods have been experimentally validated using publicly available image datasets and the results of experiments are presented in the thesis. The face description approach proposed in this thesis has been validated in several external studies, and it has been utilized and further developed by several research groups working on face analysis.