Neste trabalho é apresentada a implantação de um servidor de imagens médicas com a implementação e integração de módulos para recuperação textual e baseada em conteúdo para o Serviço de Radiodiagnóstico do Hospital das Clínicas da Faculdade de Medicina de Ribeirão Preto (HCFMRP) da Universidade de São Paulo (USP). O sistema permite a aquisição, gerenciamento, armazenamento e disponibilização das informações dos pacientes, seus exames, laudos e imagens através da internet. Os exames radiológicos e suas respectivas imagens podem ser recuperados por informações textuais ou por similaridade do conteúdo pictório das imagens. As imagens utilizadas são de ressonância magnética nuclear e tomografia computadorizada e são geradas no padrão DICOM 3.0. O sistema foi desenvolvido contemplando tecnologias para Web com interfaces amigáveis para recuperação das informações. Ele é composto por três módulos integrados, sendo o servidor de imagens, o módulo de consulta textual e o módulo de consulta por similaridade. Os resultados apresentados indicam que as imagens são gerenciadas e armazenadas corretamente, bem como o tempo de retorno das imagens é clinicamente satisfatório, tanto para a consulta textual como para a consulta por similaridade. As avaliações da recuperação por similaridade apresentam que o extrator escolhido pode ser considerado relevante para separar as imagens por região anatômica. / This work introduces an the development of a server of medical images with the implementation and integration of modules to query/retrieve text information and content-based to Radiology Service of Hospital das Clínicas da Faculdade de Medicina de Ribeirão Preto (HCFMRP) at Universidade de São Paulo (USP). The system allows the acquisition, management, archiving and availability of the patients information, theirs exams, results and images through of internet. The radiological exams and theirs respectives images can be retrieved by text information or similarity of pictorial content of images. Images are from magnetic resonance nuclear and computadorized tomography and are given using DICOM 3.0 protocol. The system has been developed considering web technologies with friendly interfaces to retrieval of information. It is composed by three integrated modules: the image server module, the query text module and query by similarity module. Results show that images are managed and archived exactly, retrieval time of images is clinically satisfactory, considering both the text query as well as the query by similarity. The evaluation of the retrieval by similarity shows the chosen extractor can be considerated relevant to separate the images by anatomic region.

Made available in DSpace on 2016-06-02T19:06:00Z (GMT). No. of bitstreams: 1 4756.pdf: 1515606 bytes, checksum: 12146689055c9826f258e527c3ae001a (MD5) Previous issue date: 2012-11-19 / Universidade Federal de Sao Carlos / This work proposes two new techniques of feature vector pre-processing to improve CBIR and image classification systems: a method of feature transformation based on the k-means clustering approach (Feature Transformation based on K-means - FTK) and a method of Weighted Feature Discretization - WFD. The FTK method employs the clustering principle of k-means to compact the feature vector space. The WFD method performs a weighted feature discretization, privileging the most important feature ranges to distinguish images. The proposed methods were employed to pre-process the feature vector in CBIR and in classification approaches, comparing the results with the pre-processing performed by PCA (a well known feature transformation method) and the original feature vector: FTK produced a reduction in the feature vector size with an improving in the query precision and a improvement in the classification accuracy; WFD improved the query precision up to and a improvement in the classification accuracy; the combination of WFD and FTK improved also the query precision and a improvement in the classification accuracy. These are very important results, especially when compared with PCA results, which leads to a minor reduction in the feature vector size, a minor increase in the query precision and a minor increase in the classification accuracy. Also the proposed approaches have linear computational cost where PCA has a cubic computational cost. The results indicate that the proposed approaches are well-suited to perform image feature vector pre-processing improving the overall quality of CBIR and classification systems. / Neste trabalho, propomos diminuir o gap semântico e os problemas de maldição de dimensionalidade apresentando duas técnicas de préprocessamento do vetor de características com o objetivo de melhorar a recuperação de imagens baseada em conteúdo e sistemas de classificação de imagens: um método de redução de dimensionalidade do vetor de características original, baseado no algoritmo k-means, chamado FTK (Feature Transformation based on K-means) e um método de discretização ponderada de características que privilegia as faixas de características mais importantes para distinguir imagens, chamado WFD (Weighted Feature Discretization). Os métodos propostos foram utilizados para pré-processar os vetores de características nas abordagens CBIR e classificação, comparando o pré-processamento executado pelo método PCA e os resultados dos vetores de características originais. O algoritmo FTK promoveu uma redução no tamanho do vetor de características com uma melhoria na precisão da consulta e na precisão de classificação. O algoritmo WFD melhorou a precisão da consulta e classificação; a combinação de dos dois algoritmos propostos também melhorou a precisão da consulta e classificação. Estes resultados são muito importantes, especialmente quando comparados com os resultados do método PCA, que também leva a uma redução no tamanho do vetor de características, a um menor aumento na precisão da consulta e a menor aumento na precisão da classificação. Além disso, as técnicas propostas têm custo computacional linear, enquanto o PCA tem um custo computacional cúbico. Os resultados indicam que os métodos propostos são abordagens adequadas para realizar pré-processamento dos vetores de características de imagens em sistemas CBIR e em sistemas de classificação.

Dans le cadre du projet Européen X-Media, de nombreuses contributions ont été apportées aux problèmes de classification d'image et de recherche d'images par le contenu dans des contextes industriels hétérogènes. Ainsi, après avoir établi un état de l'art des descripteurs d'image les plus courant, nous nous sommes dans un premier temps intéressé a des méthodes globales, c'est à dire basée sur la description totale de l'image par des descripteurs. Puis, nous nous sommes attachés a une analyse plus fine du contenu des images afin d'en extraire des informations locales, sur la présence et la localisation d'objets d'intérêt. Enfin, nous avons proposé une méthode hybride de recherche d'image basée sur le contenu qui s'appuie sur la description locale des régions de l'image afin d'en tirer une signature pouvant être utilisée pour des requêtes globales et locales. / The explosive development of affordable, high quality image acquisition deviceshas made available a tremendous amount of digital content. Large industrial companies arein need of efficient methods to exploit this content and transform it into valuable knowledge.This PhD has been accomplished in the context of the X-MEDIA project, a large Europeanproject with two major industrial partners, FIAT for the automotive industry andRolls-Royce plc. for the aircraft industry. The project has been the trigger for research linkedwith strong industrial requirements. Although those user requirements can be very specific,they covered more generic research topics. Hence, we bring several contributions in thegeneral context of Content-Based Image Retrieval (CBIR), Indexing and Classification.In the first part of the manuscript we propose contributions based on the extraction ofglobal image descriptors. We rely on well known descriptors from the literature to proposemodels for the indexing of image databases, and the approximation of a user defined categorisation.Additionally, we propose a new descriptor for a CBIR system which has toprocess a very specific image modality, for which traditional descriptors are irrelevant. Inthe second part of the manuscript, we focus on the task of image classification. Industrialrequirements on this topic go beyond the task of global image classification. We developedtwo methods to localize and classify the local content of images, i.e. image regions, usingsupervised machine learning algorithms (Support Vector Machines). In the last part of themanuscript, we propose a model for Content-Based Image Retrieval based on the constructionof a visual dictionary of image regions. We extensively experiment the model in orderto identify the most influential parameters in the retrieval efficiency.

The aim of this master thesis is to evaluate the performance of the content-based image retrieval system ImBrowse from a semantic point of view. Evaluation of retrieval performance is a problem in content-based image retrieval (CBIR). There are many different methods for measuring the performance of content-based image retrieval systems, but no common way for performing the evaluation. The main focus is on image retrieval regarding the extraction of the visual features in the image, from three semantic levels. The thesis tries to elucidate the semantic gap, which is the problem when the systems extraction of the visual features from the image and the user’s interpretation of that same information do not correspond. The method of this thesis is based on similar methods in evaluation studies of CBIR systems. The thesis is an evaluation of ImBrowse’s feature descriptors for 30 topics at three semantic levels and compared the descriptors performance based on our relevance assessment. For the computation of the results the precision at DCV = 20 is used. The results are presented in tables and a chart. The conclusion from this evaluation is that the retrieval effectiveness from a general point of view did not meet the semantic level of our relevance assessed topics. However, since the thesis do not have another system with the same search functions to compare with it is difficult to draw a comprehensive conclusion of the results. / Uppsatsnivå: D

As the information society becomes increasingly flooded with digital images, the need for efficient image retrieval systems increases as well. To handle the vast amounts of data involved, the indexing process needs to be run automatically, using content-based descriptors extracted directly from the digital image, such as colour composition, shape and texture features. These content-based image retrieval systems are often slow and cumbersome, and can appear confusing to an ordinary user who does not understand the underlying mechanisms. One step towards more efficient and user-friendly retrieval systems might be to adjust the weight placed on various descriptors depending on which image category is being searched for. The results of this thesis show that certain categories of digital images would benefit from having extra weight assigned to colour, texture or shape features when searching for images of that category.

Ce travail porte sur l'indexation automatique de photographies personnelles par des concepts visuels de haut niveau d'abstraction. Nous argumentons en faveur d'une approche basée sur l'apprentissage non supervisé, en mettant en avant les limites de l'apprentissage supervisé. Nous proposons un paradigme d'apprentissage non supervisé basé sur deux types de régularités, correspondant respectivement aux notions de structure et de similarité. Ces régularités sont apprises à partir d'un flux d'informations visuelles et constituent les nœuds d'un réseau grandissant. Les données d'apprentissage sont recodées en termes des connaissances déjà acquises. Des expérimentations sur des données réelles et synthétisées montrent que notre approche permet de créer une représentation des données pertinente, engendrant une indexation de meilleure qualité. Ces expérimentations très prometteuses permettent d'esquisser des perspectives ambitieuses.

Submitted by Júllia Mendes Pestana Dos Santos (julliapestan@hotmail.com) on 2018-07-16T01:31:34Z No. of bitstreams: 1 DISSERTAÇÃO FINAL.output.pdf: 3255322 bytes, checksum: 56ad1a6301c009bcfaf0bd6c5ef7b0f4 (MD5) / Approved for entry into archive by Telma Jaqueline Dias Silveira null (telmasbl@marilia.unesp.br) on 2018-07-16T19:46:08Z (GMT) No. of bitstreams: 1 santos_jmp_me_mar.pdf: 3255322 bytes, checksum: 56ad1a6301c009bcfaf0bd6c5ef7b0f4 (MD5) / Made available in DSpace on 2018-07-16T19:46:08Z (GMT). No. of bitstreams: 1 santos_jmp_me_mar.pdf: 3255322 bytes, checksum: 56ad1a6301c009bcfaf0bd6c5ef7b0f4 (MD5) Previous issue date: 2018-06-18 / A questão da indexação e recuperação de imagens tem atraído a atenção de novos interessados, tanto os preocupados com o conteúdo informacional, quanto os interessados em desenvolver formas automatizadas de descrição e acesso ao conteúdo imagético. No contexto das imagens digitais, para que elas se tornem recursos disponíveis, recuperáveis e acessíveis é necessário um desenvolvimento de processos para construção de representações. Um impulso na revolução das imagens digitais foi dado pela expansão da utilização dos computadores, onde surgiram técnicas para captura, armazenamento, processamento e transmissão das mesmas. Neste sentido, o objetivo do trabalho foi discutir o processo de indexação e recuperação de imagens no contexto de sistemas informáticos. Abordamos tanto a indexação quanto a recuperação de imagens fotográficas, visando melhorar a precisão de recuperação desses documentos em sistemas de informação. O uso de metadados neste contexto é significativo, pois o funcionamento da web está ligado à facilidade de recuperação de dados. E assim, os sistemas de Recuperação da Imagem Baseado em Conteúdo (CBIR) evidenciam a necessidade da introdução de novos atributos/características, como a utilização de cor, formas, texturas. Mas somente essas propriedades não solucionam o problema da organização e recuperação de conteúdo imagético e, portanto nesse contexto a criação de metadados aos elementos básicos da linguagem visual: ponto, linha, forma, direção, tom, cor, textura, escala, dimensão e movimento para construção e recuperação da imagem serão a proposta e análise dessa pesquisa. E os resultados apontam que o software Sepiades por ser um software livre, ainda que seja para descrição de imagens, torna-se mais acessível e compatível e seus elementos são passíveis de miscigenação, sendo uma boa solução. Conclui-se que uma participação significativa dos indexadores no desenvolvimento de softwares para organização e recuperação imagética é de extrema importância e contribuição para área da Ciência da Informação (CI) e Computação. Desse modo destacamos a importância de estudos e pesquisas que tenham como objetivo a Recuperação de Imagens e os diversos mecanismos de busca da informação, bem como, conhecer as tecnologias existentes e a sua utilização. / The issue of indexing and retrieval of images has attracted the attention of new stakeholders, both those concerned with the information content of the images and those interested in developing automated forms of description and access to images. In the context of digital images, for them to become available, retrievable and accessible resources, a process development is necessary for the construction of representations. An impulse in the digital imaging revolution was the expansion of computer use, where techniques for capturing, storing, processing and transmitting images emerged. In this sense, the objective of the work is the process of indexing and retrieving images in the context of computer systems. We address both indexing and retrieval of photographic images, aiming to improve the accuracy of image retrieval in information systems. The use of metadata in this context will be significant, as the operation of the web is linked to the ease of data recovery. Thus, Content Based Image Retrieval (CBIR) systems highlight the need to introduce new attributes / characteristics, such as the use of color, shapes, textures. But only these properties do not solve the problem of image organization and retrieval, and in this context, the creation of metadata to the basic elements of visual language: point, line, shape, direction, tone, color, texture, scale, dimension and motion for construction and image recovery will be the proposal and analysis of this research. And the results show that the software Sepiades, because it is a free software, although it is for image description, it becomes more accessible and compatible and its elements are susceptible of miscegenation, being a good solution. It is concluded that a significant participation of indexers in the development of software for organization and image recovery is of extreme importance and contribution to the area of Information Science (CI) and Computing. In this way, we highlight the importance of studies and research that aim at Image Retrieval and in the various information search mechanisms, as well as to know the existing technologies and their use.

Submitted by Bruno Tavares Padovese null (bpadovese@gmail.com) on 2017-07-03T15:22:41Z No. of bitstreams: 1 Dissertacao_Mestrado_Bruno_Tavares_Padovese.pdf: 4559390 bytes, checksum: 9152719c817205d08d3a72b5a5abc949 (MD5) / Approved for entry into archive by Luiz Galeffi (luizgaleffi@gmail.com) on 2017-07-04T17:59:03Z (GMT) No. of bitstreams: 1 padovese_bt_me_sjrp.pdf: 4559390 bytes, checksum: 9152719c817205d08d3a72b5a5abc949 (MD5) / Made available in DSpace on 2017-07-04T17:59:03Z (GMT). No. of bitstreams: 1 padovese_bt_me_sjrp.pdf: 4559390 bytes, checksum: 9152719c817205d08d3a72b5a5abc949 (MD5) Previous issue date: 2017-05-15 / Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) / Resumo: Os estágios iniciais da doença de Alzheimer são comumente confundidos com o processo natural de envelhecimento. Adicionalmente, a metodologia envolvida no diagnóstico por radiologistas pode ser subjetiva e difícil de documentar. Neste cenário, o desenvolvimento de abordagens acessíveis capazes de auxiliar no diagnóstico precoce da doença de Alzheimer é crucial. Várias abordagens têm sido empregadas com este objetivo, especialmente utilizando imagens de ressonância magnética cerebral. Embora resultados com precisão satisfatória tenham sido obtidos, a maioria das abordagens requer etapas de pré-processamento muito específicas, baseadas na anatomia do cérebro. Neste trabalho, apresentamos uma nova abordagem de recuperação de imagens para auxílio ao diagnóstico da doença de Alzheimer, com base em descritores de propósito geral e uma etapa de pós-processamento não supervisionada. Os exames de ressonância magnética cerebral são processados e recuperados através de descritores de uso geral sem nenhuma etapa de pré-processamento. Dois algoritmos de aprendizado não-supervisionados baseados em ranqueamento foram aplicados para melhorar a eficácia dos resultados iniciais: os algoritmos RL-Sim e ReckNN. Os resultados experimentais demonstram que a abordagem proposta é capaz de atingir resultados de recuperação eficazes, sendo adequada para auxiliar no diagnóstico da doença de Alzheimer. / Abstract: Initial stages of Alzheimer’s disease are easily confused with the normal aging process. Additionally, the methodology involved in the diagnosis by radiologists can be subjective and difficult to document. In this scenario, the development of accessible approaches capable of supporting the early diagnosis of Alzheimer’s disease is crucial. Various approaches have been employed with this objective, specially using brain MRI scans. Although certain satisfactory accuracy results have been achieved, most of the approaches require very specific pre-processing steps based on the brain anatomy. In this work, we present a novel image retrieval approach for supporting the Alzheimer’s disease diagnostic, based on general purpose features and an unsupervised post-processing step. The brain MRI scans are processed and retrieved through general visual features without any pre-processing step. Two rank-based unsupervised distance learning algorithms were used for improving the effectiveness of the initial results: the RL-Sim and ReckNN algorithms. Experimental results demonstrate that the proposed approach can achieve effective retrieval results, being suitable in aiding the diagnosis of Alzheimer’s disease. / CNPq: 154034/2016-9

Information retrieval systems are playing an important role in our day to day life for getting the required information. Many text retrieval systems are available and are working successfully. Even though internet is full of other media like images, audio and video, retrieval systems for these media are rare and have not achieved success as that of text retrieval systems. Image retrieval systems are useful in many applications; there is a high demand for effective and efficient tool for image organization and retrieval as per users need. Images are classified into text based image retrieval and content based image retrieval, we proposed a text based image retrieval system, which makes use of ontology to make the retrieval process intelligent. We worked on Cricket World Cup 2007. We combined text based image retrieval approach with content based image retrieval, which uses color and texture as basic low level features. / kvishu223@gmail.com, pravs72@yahoo.co.in.

Online shopping is becoming a popular option for consumers. Currently, the most common product searching method that online shopping websites provide is keyword search. Most shoppers have to carefully select relevant keywords to search for their favorite products. Finding desired products using a query image for online shopping is currently not available. Image has been used for searching similar images in the database but they are usually not well annotated. Research effort has been devoted to developing reliable image-based retrieval systems for applications such as medical image retrieval and trademark search. None of these developments focuses on improving online shopping experiences for consumers. This thesis reports the development of an image retrieval system to provide better online shopping experience for consumers. The system searches products with similar appearance such as shape and textures to the query images the user provides. Turn angle is a contour based shape descriptor. It has many unique properties that make it a perfect shape matching method for image retrieval. The best matching image has the shortest shape distance to the query shape. Turn angle, however, could fail with slightly stretched shapes. Dynamic programming is used to help turn angle match slightly deformed shapes. Another technique called centroid distance is also included as a restriction for shape matching in order to avoid retrieving irrelevant or disparate shapes. With a well-built database, the enhanced turn angle descriptor that includes dynamic programming and centroid distance is able to reach a high accuracy rate.Shape matching alone is usually not sufficient for a powerful retrieval system. Products with similar shape but very different textures will not be distinguished based solely on shape matching. Edge histogram is a robust shape descriptor for texture matching. It can be implemented to construct either global or local histogram for this purpose. Global edge histogram uses only 5 bins, which is simple but ignores detail texture information. Local and semi-global edge histograms are more complex but retains detail texture information. A hierarchical matching system is built to combine the shape and texture descriptors for better retrieval accuracy.Easy access to the shopping system is desired. An Android Application is developed to provide consumers a convenient and friendly tool to use the system. Grab cut is applied to the captured image to segment the object from the background. The segmentation provides the retrieval system the required contour information for shape matching. The Android Application submits the captured image along with the segmented contour to the server. After the retrieval process is completed, the server sends retrieved images of similar products back to the Android App for the user to consider. Using the retrieval system via a handheld device provides a user-friendly online shopping experience.