The retinal vasculature analysis plays an important role in the diagnosis of ophthalmological diseases, as well as general human disorders that manifest on the retina. The fundus photograph is a 2-D color image modality of the retina and is widely used in modern ophthalmology clinics due to its relatively low cost and its non-invasive access to the retina. However, due to the complexity of the retinal vasculature presented on the image and the large variation of the image quality, no automated method is able to re-construct the retinal vasculature (i.e. construct arteriovenous trees) satisfactorily, thus preventing its analysis on large-scale clinical datasets. In this thesis, we present a systematic and complete study to automatically construct the retinal vasculature on fundus photographs and apply it to a clinical dataset. First of all, a preliminary study is conducted to detect and classify important landmarks in the retinal vasculature using a machine learning method. The evaluation of this method reveals the difficulty of identifying each landmark as an independent target. Then a novel and more global method is proposed to construct retinal arteriovenous trees (A/V trees). The strategy of the proposed method is to build an over-connected vessel network, and separate it into vascular trees, then classify them into A/V trees. Particularly, by taking advantages of specific properties of the retinal vasculature, global and local information are combined together to recognize landmarks of the vasculature. Instead of recognizing each landmark independently as other methods do, this method considers the relationship between landmarks in a more global manner, thus recognizing them simultaneously and globally. With a special graph design, each landmark is associated with multiple possible configurations and costs, and a near optimal solution is selected by minimizing the costs of landmarks and the global property of the whole vascular network. With each landmark recognized, the A/V trees are easily inferred with a pixel classification method. By doing so, local noise in the images and local errors during pre-processing are corrected to some degree, and small vessels that are difficult to classify locally can also be recognized. The proposed method is compared with another method and the evaluation demonstrates its superiority. To demonstrate its potential applicability, we apply the proposed method on a cohort study data of HIV-infected patients with treatment. New metrics to analyze retinal vessel width is developed based on the A/V trees built using the proposed method, and it is compared with a conventional metric. Statistical analysis reveals the advantages of the new metric and thus indicates the benefit of the proposed method and its potential application on large datasets.

Cette thèse a pour objet l’étude de nouvelles méthodes de traitement d’image appliquées à l’analyse d’images numériques du fond d'œil de patients diabétiques. En particulier, nous nous sommes concentrés sur le développement algorithmique supportant un système de dépistage automatique de la rétinopathie diabétique. Les techniques présentées dans ce document peuvent être classées en trois catégories: (1) l’évaluation et l’amélioration de la qualité d’image, (2) la segmentation des lésions, et (3) le diagnostic. Pour la première catégorie, nous présentons un algorithme rapide permettant l’estimation numérique de la qualité d’une seule image à partir de caractéristiques extraites de la vascularisation et de la couleur du fond d'œil. De plus, nous démontrons qu’il est possible d’augmenter la qualité des images et de supprimer les artefacts de réflexion en fusionnant les informations extraites de plusieurs images d’un même fond d'œil (images capturées en changeant le point d’attention regardé par le patient). Pour la deuxième catégorie, deux familles de lésion sont ciblées: les exsudats et les microanévrysmes. Deux nouveaux algorithmes pour l’analyse des images du fond d'œil sont proposés et comparés avec les techniques existantes afin de démontrer leur efficacité. Dans le cas des microanévrismes, une nouvelle méthode basée sur la transformée de Radon a été développée. Dans la dernière catégorie, nous présentons un algorithme permettant de diagnostiquer la rétinopathie diabétique et les œdèmes maculaires en analysant les lésions détectées par segmentation d’image; à partir d’une seule image, notre algorithme permet de diagnostiquer une rétinopathie diabétique et/ou un œdème maculaire en ~ 22 secondes sur une machine à 1,6 GHz avec 4 Go de RAM; de plus, nous montrons les premiers résultats de notre algorithme de détection d'œdème maculaire basé sur des images du fond d'œil multiples, qui peut éventuellement permettre d’identifier le gonflement de la macula même si aucune lésion n’est visible. / In this Ph.D. thesis, we study new methods to analyse digital fundus images of diabetic patients. In particular, we concentrate on the development of the algorithmic components of an automatic screening system for diabetic retinopathy. The techniques developed can be categorized in: quality assessment and improvement, lesion segmentation and diagnosis. For the first category, we present a fast algorithm to numerically estimate the quality of a single image by employing vasculature and colour-based features; additionally, we show how it is possible to increase the image quality and remove reflection artefacts by merging information gathered in multiple fundus images (which are captured by changing the stare point of the patient). For the second category, two families of lesion are targeted: exudate and microaneurysms; two new algorithms which work on single fundus images are proposed and compared with existing techniques in order to prove their efficacy; in the microaneurysms case, a new Radon transform-based operator was developed. In the last diagnosis category, we have developed an algorithm that diagnoses diabetic retinopathy and diabetic macular edema based on the lesions segmented; starting from a single unseen image, our algorithm can generate a diabetic retinopathy and ma cular edema diagnosis in _22 seconds on a 1.6 GHz machine with 4 GB of RAM; additionally, we show the first results of a macular edema detection algorithm based on multiple fundus images, which can potentially identify the swelling of the macula even when no lesions are visible.

Retinógrafo é um complexo sistema óptico que, simultaneamente, ilumina e captura imagens da retina. Basicamente, esse equipamento é composto por três módulos: iluminação, lente objetiva e detecção. Em razão de seu sofisticado desenho óptico, é possível conciliar baixa refletividade da retina e obtenção de imagens de alta qualidade. Entretanto, em virtude do alto número de componentes não-coaxiais do módulo de iluminação, seu alinhamento óptico se torna complexo. Neste trabalho, é apresentado um sistema óptico totalmente coaxial para um retinógrafo nãomidriático. A óptica de iluminação tradicional é substituída por um anel de Surface Mounted Device (SMD) Light Emitting Diodes (LEDs) que, analogamente ao equipamento tradicional, forma um anel de luz no plano da pupila do olho, iluminando homogeneamente a retina e evitando reflexos gerados na córnea. Com essa substituição, os três módulos do equipamento se tornam coaxiais, facilitando o alinhamento final. Devido à inovação da arquitetura do retinógrafo, um novo método de eliminação de reflexos também foi introduzido, possibilitando ao equipamento fornecer imagens nítidas e de alta qualidade, suficientes para exames de triagem. Além da iluminação, os módulos da objetiva e de detecção foram substituídos por componentes comerciais, visando a simplificação do projeto de retinógrafo. Dessa forma, pretende-se reduzir o custo de comercialização do produto, de modo que clínicas no Brasil e em países em desenvolvimento possam ser equipadas e capazes de realizar diagnósticos de doenças do olho que causam perda parcial ou total da visão. / Fundus camera is a complex optical system that simultaneously illuminates and images the retina. It is basically divided into three modules: objective lens, illumination and detection. Because of its sophisticate optical design, it is possible to achieve high-quality images under low reflected light by the fundus. However, due to its high number of off-axis components, mainly in the illumination system, the optical alignment of the equipment can be complex. To simplify the architecture of the equipment, we report a completely coaxial optical system, with no off-axis components. The traditional illumination system is replaced by a ring of light emitting diodes of surface mounted device type. As in the previous design, the eye pupil is illuminated with a ring of light, producing a uniform pattern on the retina and avoiding reflection on the cornea. Due to this new design and the lack of optical components in the illumination system, a new method of avoiding reflection on the surfaces of the objective lens is presented. Besides, the objective lens and the detection system were composed of commercial components, also simplifying the equipment project and lowering its cost. The final goal of this work is to provide non-mydriatic high quality fundus images for screening with a low-cost equipment, enabling developing countries to increase the number of people examined.

Retinógrafo é um complexo sistema óptico que, simultaneamente, ilumina e captura imagens da retina. Basicamente, esse equipamento é composto por três módulos: iluminação, lente objetiva e detecção. Em razão de seu sofisticado desenho óptico, é possível conciliar baixa refletividade da retina e obtenção de imagens de alta qualidade. Entretanto, em virtude do alto número de componentes não-coaxiais do módulo de iluminação, seu alinhamento óptico se torna complexo. Neste trabalho, é apresentado um sistema óptico totalmente coaxial para um retinógrafo nãomidriático. A óptica de iluminação tradicional é substituída por um anel de Surface Mounted Device (SMD) Light Emitting Diodes (LEDs) que, analogamente ao equipamento tradicional, forma um anel de luz no plano da pupila do olho, iluminando homogeneamente a retina e evitando reflexos gerados na córnea. Com essa substituição, os três módulos do equipamento se tornam coaxiais, facilitando o alinhamento final. Devido à inovação da arquitetura do retinógrafo, um novo método de eliminação de reflexos também foi introduzido, possibilitando ao equipamento fornecer imagens nítidas e de alta qualidade, suficientes para exames de triagem. Além da iluminação, os módulos da objetiva e de detecção foram substituídos por componentes comerciais, visando a simplificação do projeto de retinógrafo. Dessa forma, pretende-se reduzir o custo de comercialização do produto, de modo que clínicas no Brasil e em países em desenvolvimento possam ser equipadas e capazes de realizar diagnósticos de doenças do olho que causam perda parcial ou total da visão. / Fundus camera is a complex optical system that simultaneously illuminates and images the retina. It is basically divided into three modules: objective lens, illumination and detection. Because of its sophisticate optical design, it is possible to achieve high-quality images under low reflected light by the fundus. However, due to its high number of off-axis components, mainly in the illumination system, the optical alignment of the equipment can be complex. To simplify the architecture of the equipment, we report a completely coaxial optical system, with no off-axis components. The traditional illumination system is replaced by a ring of light emitting diodes of surface mounted device type. As in the previous design, the eye pupil is illuminated with a ring of light, producing a uniform pattern on the retina and avoiding reflection on the cornea. Due to this new design and the lack of optical components in the illumination system, a new method of avoiding reflection on the surfaces of the objective lens is presented. Besides, the objective lens and the detection system were composed of commercial components, also simplifying the equipment project and lowering its cost. The final goal of this work is to provide non-mydriatic high quality fundus images for screening with a low-cost equipment, enabling developing countries to increase the number of people examined.

In this Ph.D. thesis, we study new methods to analyse digital fundus images of diabetic patients. In particular, we concentrate on the development of the algorithmic components of an automatic screening system for diabetic retinopathy. The techniques developed can be categorized in: quality assessment and improvement, lesion segmentation and diagnosis. For the first category, we present a fast algorithm to numerically estimate the quality of a single image by employing vasculature and colour-based features; additionally, we show how it is possible to increase the image quality and remove reflection artefacts by merging information gathered in multiple fundus images (which are captured by changing the stare point of the patient). For the second category, two families of lesion are targeted: exudate and microaneurysms; two new algorithms which work on single fundus images are proposed and compared with existing techniques in order to prove their efficacy; in the microaneurysms case, a new Radon transform-based operator was developed. In the last diagnosis category, we have developed an algorithm that diagnoses diabetic retinopathy and diabetic macular edema based on the lesions segmented; starting from a single unseen image, our algorithm can generate a diabetic retinopathy and ma cular edema diagnosis in _22 seconds on a 1.6 GHz machine with 4 GB of RAM; additionally, we show the first results of a macular edema detection algorithm based on multiple fundus images, which can potentially identify the swelling of the macula even when no lesions are visible.

Segmentation of vasculature tree is an important step of the process of image processing. There are many methods of automatic blood vessel segmentation. These methods are based on matched filters, pattern recognition or image classification. Use of automatic retinal image processing greatly simplifies and accelerates retinal images diagnosis. The aim of the automatic image segmentation algorithms is thresholding. This work primarily deals with retinal image thresholding. We discuss a few works using local and global image thresholding and supervised image classification to segmentation of blood tree from retinal images. Subsequently is to set of results from two different methods used image classification and discuss effectiveness of the vessel segmentation. Use image classification instead of global thresholding changed statistics of first method on healthy part of HRF. Sensitivity and accuracy decreased to 62,32 %, respectively 94,99 %. Specificity increased to 95,75 %. Second method achieved sensitivity 69.24 %, specificity 98.86% and 95.29 % accuracy. Combining the results of both methods achieved sensitivity up to72.48%, specificity to 98.59% and the accuracy to 95.75%. This confirmed the assumption that the classifier will achieve better results. At the same time, was shown that extend the feature vector combining the results from both methods have increased sensitivity, specificity and accuracy.

Image registration is widely used in clinical practice. However image registration and its~evaluation is still challenging especially with regards to new possibilities of various modalities. One of these areas is contrast-enhanced ultrasound imaging. The time-dependent image contrast, low signal-to-noise ratio and specific speckle pattern make preprocessing and image registration difficult. In this thesis a method for registration of images in ultrasound contrast-enhanced sequences is proposed. The method is based on automatic fragmentation into image subsequences in which the images with similar characteristics are registered. The new evaluation method based on comparison of perfusion model is proposed. Registration and evaluation method was tested on a flow phantom and real patient data and compared with a standard methods proposed i literature. The second part of this thesis contains examples of application of image registration in~ophthalmology and proposition for its improvement. In this area the image registration methods are widely used, especially landmark based image registration method. In this thesis methods for landmark detection and its correspondence estimation are proposed.

Fundus kamera je široce dostupné zobrazovací zařízení, které umožňuje relativně rychlé a nenákladné vyšetření zadního segmentu oka – sítnice. Z těchto důvodů se mnoho výzkumných pracovišť zaměřuje právě na vývoj automatických metod diagnostiky nemocí sítnice s využitím fundus fotografií. Tato dizertační práce analyzuje současný stav vědeckého poznání v oblasti diagnostiky glaukomu s využitím fundus kamery a navrhuje novou metodiku hodnocení vrstvy nervových vláken (VNV) na sítnici pomocí texturní analýzy. Spolu s touto metodikou je navržena metoda segmentace cévního řečiště sítnice, jakožto další hodnotný příspěvek k současnému stavu řešené problematiky. Segmentace cévního řečiště rovněž slouží jako nezbytný krok předcházející analýzu VNV. Vedle toho práce publikuje novou volně dostupnou databázi snímků sítnice se zlatými standardy pro účely hodnocení automatických metod segmentace cévního řečiště.

En los ultimos años el número de casos de ceguera se ha reducido significativamente. A pesar de este hecho, la Organización Mundial de la Salud estima que un 80% de los casos de pérdida de visión (285 millones en 2010) pueden ser evitados si se diagnostican en sus estadios más tempranos y son tratados de forma efectiva. Para cumplir esta propuesta se pretende que los servicios de atención primaria incluyan un seguimiento oftalmológico de sus pacientes así como fomentar campañas de cribado en centros proclives a reunir personas de alto riesgo. Sin embargo, estas soluciones exigen una alta carga de trabajo de personal experto entrenado en el análisis de los patrones anómalos propios de cada enfermedad. Por lo tanto, el desarrollo de algoritmos para la creación de sistemas de cribado automáticos juga un papel vital en este campo. La presente tesis persigue la identificacion automática del daño retiniano provocado por dos de las patologías más comunes en la sociedad actual: la retinopatía diabética (RD) y la degenaración macular asociada a la edad (DMAE). Concretamente, el objetivo final de este trabajo es el desarrollo de métodos novedosos basados en la extracción de características de la imagen de fondo de ojo y clasificación para discernir entre tejido sano y patológico. Además, en este documento se proponen algoritmos de pre-procesado con el objetivo de normalizar la alta variabilidad existente en las bases de datos publicas de imagen de fondo de ojo y eliminar la contribución de ciertas estructuras retinianas que afectan negativamente en la detección del daño retiniano. A diferencia de la mayoría de los trabajos existentes en el estado del arte sobre detección de patologías en imagen de fondo de ojo, los métodos propuestos a lo largo de este manuscrito evitan la necesidad de segmentación de las lesiones o la generación de un mapa de candidatos antes de la fase de clasificación. En este trabajo, Local binary patterns, perfiles granulométricos y la dimensión fractal se aplican de manera local para extraer información de textura, morfología y tortuosidad de la imagen de fondo de ojo. Posteriormente, esta información se combina de diversos modos formando vectores de características con los que se entrenan avanzados métodos de clasificación formulados para discriminar de manera óptima entre exudados, microaneurismas, hemorragias y tejido sano. Mediante diversos experimentos, se valida la habilidad del sistema propuesto para identificar los signos más comunes de la RD y DMAE. Para ello se emplean bases de datos públicas con un alto grado de variabilidad sin exlcuir ninguna imagen. Además, la presente tesis también cubre aspectos básicos del paradigma de deep learning. Concretamente, se presenta un novedoso método basado en redes neuronales convolucionales (CNNs). La técnica de transferencia de conocimiento se aplica mediante el fine-tuning de las arquitecturas de CNNs más importantes en el estado del arte. La detección y localización de exudados mediante redes neuronales se lleva a cabo en los dos últimos experimentos de esta tesis doctoral. Cabe destacar que los resultados obtenidos mediante la extracción de características "manual" y posterior clasificación se comparan de forma objetiva con las predicciones obtenidas por el mejor modelo basado en CNNs. Los prometedores resultados obtenidos en esta tesis y el bajo coste y portabilidad de las cámaras de adquisión de imagen de retina podrían facilitar la incorporación de los algoritmos desarrollados en este trabajo en un sistema de cribado automático que ayude a los especialistas en la detección de patrones anomálos característicos de las dos enfermedades bajo estudio: RD y DMAE. / In last years, the number of blindness cases has been significantly reduced. Despite this promising news, the World Health Organisation estimates that 80% of visual impairment (285 million cases in 2010) could be avoided if diagnosed and treated early. To accomplish this purpose, eye care services need to be established in primary health and screening campaigns should be a common task in centres with people at risk. However, these solutions entail a high workload for trained experts in the analysis of the anomalous patterns of each eye disease. Therefore, the development of algorithms for automatic screening system plays a vital role in this field. This thesis focuses on the automatic identification of the retinal damage provoked by two of the most common pathologies in the current society: diabetic retinopathy (DR) and age-related macular degeneration (AMD). Specifically, the final goal of this work is to develop novel methods, based on fundus image description and classification, to characterise the healthy and abnormal tissue in the retina background. In addition, pre-processing algorithms are proposed with the aim of normalising the high variability of fundus images and removing the contribution of some retinal structures that could hinder in the retinal damage detection. In contrast to the most of the state-of-the-art works in damage detection using fundus images, the methods proposed throughout this manuscript avoid the necessity of lesion segmentation or the candidate map generation before the classification stage. Local binary patterns, granulometric profiles and fractal dimension are locally computed to extract texture, morphological and roughness information from retinal images. Different combinations of this information feed advanced classification algorithms formulated to optimally discriminate exudates, microaneurysms, haemorrhages and healthy tissues. Through several experiments, the ability of the proposed system to identify DR and AMD signs is validated using different public databases with a large degree of variability and without image exclusion. Moreover, this thesis covers the basics of the deep learning paradigm. In particular, a novel approach based on convolutional neural networks is explored. The transfer learning technique is applied to fine-tune the most important state-of-the-art CNN architectures. Exudate detection and localisation tasks using neural networks are carried out in the last two experiments of this thesis. An objective comparison between the hand-crafted feature extraction and classification process and the prediction models based on CNNs is established. The promising results of this PhD thesis and the affordable cost and portability of retinal cameras could facilitate the further incorporation of the developed algorithms in a computer-aided diagnosis (CAD) system to help specialists in the accurate detection of anomalous patterns characteristic of the two diseases under study: DR and AMD. / En els últims anys el nombre de casos de ceguera s'ha reduït significativament. A pesar d'este fet, l'Organització Mundial de la Salut estima que un 80% dels casos de pèrdua de visió (285 milions en 2010) poden ser evitats si es diagnostiquen en els seus estadis més primerencs i són tractats de forma efectiva. Per a complir esta proposta es pretén que els servicis d'atenció primària incloguen un seguiment oftalmològic dels seus pacients així com fomentar campanyes de garbellament en centres regentats per persones d'alt risc. No obstant això, estes solucions exigixen una alta càrrega de treball de personal expert entrenat en l'anàlisi dels patrons anòmals propis de cada malaltia. Per tant, el desenrotllament d'algoritmes per a la creació de sistemes de garbellament automàtics juga un paper vital en este camp. La present tesi perseguix la identificació automàtica del dany retiniano provocat per dos de les patologies més comunes en la societat actual: la retinopatia diabètica (RD) i la degenaración macular associada a l'edat (DMAE) . Concretament, l'objectiu final d'este treball és el desenrotllament de mètodes novedodos basats en l'extracció de característiques de la imatge de fons d'ull i classificació per a discernir entre teixit sa i patològic. A més, en este document es proposen algoritmes de pre- processat amb l'objectiu de normalitzar l'alta variabilitat existent en les bases de dades publiques d'imatge de fons d'ull i eliminar la contribució de certes estructures retinianas que afecten negativament en la detecció del dany retiniano. A diferència de la majoria dels treballs existents en l'estat de l'art sobre detecció de patologies en imatge de fons d'ull, els mètodes proposats al llarg d'este manuscrit eviten la necessitat de segmentació de les lesions o la generació d'un mapa de candidats abans de la fase de classificació. En este treball, Local binary patterns, perfils granulometrics i la dimensió fractal s'apliquen de manera local per a extraure informació de textura, morfologia i tortuositat de la imatge de fons d'ull. Posteriorment, esta informació es combina de diversos modes formant vectors de característiques amb els que s'entrenen avançats mètodes de classificació formulats per a discriminar de manera òptima entre exsudats, microaneurismes, hemorràgies i teixit sa. Per mitjà de diversos experiments, es valida l'habilitat del sistema proposat per a identificar els signes més comuns de la RD i DMAE. Per a això s'empren bases de dades públiques amb un alt grau de variabilitat sense exlcuir cap imatge. A més, la present tesi també cobrix aspectes bàsics del paradigma de deep learning. Concretament, es presenta un nou mètode basat en xarxes neuronals convolucionales (CNNs) . La tècnica de transferencia de coneixement s'aplica per mitjà del fine-tuning de les arquitectures de CNNs més importants en l'estat de l'art. La detecció i localització d'exudats per mitjà de xarxes neuronals es du a terme en els dos últims experiments d'esta tesi doctoral. Cal destacar que els resultats obtinguts per mitjà de l'extracció de característiques "manual" i posterior classificació es comparen de forma objectiva amb les prediccions obtingudes pel millor model basat en CNNs. Els prometedors resultats obtinguts en esta tesi i el baix cost i portabilitat de les cambres d'adquisión d'imatge de retina podrien facilitar la incorporació dels algoritmes desenrotllats en este treball en un sistema de garbellament automàtic que ajude als especialistes en la detecció de patrons anomálos característics de les dos malalties baix estudi: RD i DMAE. / Colomer Granero, A. (2018). Fundus image analysis for automatic screening of ophthalmic pathologies [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/99745