Powered by OpenAIRE graph
Found an issue? Give us feedback
addClaim

This Research product is the result of merged Research products in OpenAIRE.

You have already added 0 works in your ORCID record related to the merged Research product.

CAD para mamografia

Authors: Costa, Samuel Fontes da;

CAD para mamografia

Abstract

O cancro da mama é actualmente uma das principais causas de morte entre as mulheres. Os sistemas de diagnóstico auxiliado por computador permitem o apoio aos radiologistas, com o objectivo de melhorar a precisão do diagnóstico, fornecendo uma \segunda opinião". Este trabalho apresenta uma metodologia para classificação de nódulos mamários com base em técnicas de processamento de imagens digitais para a extracção de características de textura e forma. As imagens mamográficas digitais utilizadas foram obtidas a partir da base de dados de domínio público miniMIAS (do inglês, Mammographic Image Analysis Society). Esta base de dados é constítuida por um total de 322 mamogramas, de onde foram recortados 67 regiões de interesse que continham massas. O contorno das massas foram interactivamente delineados, com o suporte do algoritmo live-wire. Assim, foram extraídas cinco características de forma baseadas no contorno e oito características de textura baseadas na matriz de co-ocorrência entre níveis de cinza, com o objectivo de caracterizar as imagens. É proposta a aplicação e estudo de redes neuronais artificiais, de modo a classificar as massas em benignas e malignas. Assim sendo, foram testadas várias redes neuronais, variando os conjuntos de entrada e as topologias das redes de modo a obter a melhor classificação. Os resultados foram comparados em termos da área Az sob a curva ROC. ABSTRACT: Breast cancer is currently one of the leading causes of death among women. Systems designed for computer-aided diagnosis provide support to radiologists, with the aim of improving the accuracy of diagnosis by providing a \second opinion". This thesis presents a methodology for classification of breast nodules, based on techniques for processing digital images developed for the extraction of texture and shape features. The digital mammographic images were obtained from a public domain database, \miniMIAS". This database as a total of 322 mammograms, from where 67 interest zones that contained masses were selected. The contours of these masses were interactively outlined with support from a \live-wire"algorithm. Then five shape features were computed as well as eight texture features based on the co-occurrence matrix. We assessed the application of artificial neural networks, in order to classify the masses in benign or malignant. Several neural networks topologies and learning algorithms were evaluated in order to obtain the best classify performance. The results were then submitted to a comparison in terms of the Az area and under the ROC curve.

Mestrado em Engenharia Electrónica e Telecomunicações

Related Organizations
Keywords

Processamento de imagem, Diagnóstico por imagem, Visualização tridimensional, Engenharia electrónica, Mamografia

  • BIP!
    Impact byBIP!
    selected citations
    These citations are derived from selected sources.
    This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    0
    popularity
    This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
    Average
    influence
    This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    Average
    impulse
    This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
    Average
Powered by OpenAIRE graph
Found an issue? Give us feedback
selected citations
These citations are derived from selected sources.
This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Citations provided by BIP!
popularity
This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
BIP!Popularity provided by BIP!
influence
This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Influence provided by BIP!
impulse
This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
BIP!Impulse provided by BIP!
0
Average
Average
Average
Related to Research communities
Cancer Research
Upload OA version
Are you the author of this publication? Upload your Open Access version to Zenodo!
It’s fast and easy, just two clicks!