Elsevier

Computers in Biology and Medicine

Volume 57, 1 February 2015, Pages 42-53
Computers in Biology and Medicine

Classification of breast regions as mass and non-mass based on digital mammograms using taxonomic indexes and SVM

https://doi.org/10.1016/j.compbiomed.2014.11.016Get rights and content
Under an Elsevier user license
open access

Highlights

  • This work proposes a methodology for classification of regions of mass and non-mass.

  • Taxonomic diversity and distinctness are used to describe the texture.

  • The tests were carried out over a sample of 3404 regions of mass and non-mass.

  • The proposed work achieved good results for the classification with accuracy of 98.8%.

Abstract

Breast cancer is the second most common type of cancer in the world. Several computer-aided detection and diagnosis systems have been used to assist health experts identify suspicious areas that are difficult to perceive with the human eye, thus aiding in the detection and diagnosis of cancer. This work proposes a methodology for the discrimination and classification of regions extracted from mammograms as mass and non-mass. The Digital Database for Screening Mammography (DDSM) was used in this work for the acquisition of mammograms. The taxonomic diversity index (Δ) and the taxonomic distinctness (Δ), which were originally used in ecology, were used to describe the texture of the regions of interest. These indexes were computed based on phylogenetic trees, which were applied to describe the patterns in regions of breast images. Two approaches were used for the analysis of texture: internal and external masks. A support vector machine was used to classify the regions as mass and non-mass. The proposed methodology successfully classified the masses and non-masses, with an average accuracy of 98.88%.

Keywords

Medical image
Breast cancer
Phylogenetic trees
Taxonomic diversity index (Δ)
Taxonomic distinctness (Δ)

Cited by (0)

Fernando Soares Sérvulo de Oliveira received the Master degree in Computer Science at Federal University of Maranhão, Brazil in 2013. His major interest nowadays is obtaining a Doctor degree. His research interests include medical image processing, pattern recognition, artificial intelligence and machine learning.

Antonio Oseas de Carvalho Filho received the Master degree in Computer Science at Federal University of Maranhão, Brazil in 2013. He is currently a Professor at the Informatics, Federal University of Maranhão, Brazil. His major interest nowadays is obtaining a Doctor degree. His research interests include medical image processing, pattern recognition, artificial intelligence, machine learning and programming WEB.

Aristófanes Corrêa Silva received a Ph.D. degree in Informatics from Pontiphical Catholic University of Rio de Janeiro, Brazil in 2004. Currently he is a Professor at the Federal University of Maranh ao (UFMA), Brazil. He teaches image processing, pattern recognition and programming language. His research interests include image processing, image understanding, medical image processing, machine vision, artificial intelligence, pattern recognition and, machine learning.

Anselmo Cardoso de Paiva received B.Sc. in civil engineering from Maranhão State Univeristy, Brazil in 1990, an M.Sc. in civil engineering-Structures and a Ph.D. in Informatics from Pontiphical Catholic University of Rio de Janeiro, Brazil, in 1993 and 2002, respectively. He is currently a Professor at the Informatics department, Federal University of Maranh ao -Brazil. His current interests include medical image processing, geographical information systems and scientific visualization.

Marcelo Gattass took his Ph.D. in 1982 from Cornell University and is a full professor at PUC-Rio׳s Computer Science Department. He currently teaches Computer Graphics and 3D Computer Vision for graduate students and Data Structures for undergraduates. He is also the Director of Tecgraf/PUC-Rio – Computer Graphics Technology Laboratory, where he supervises several industry-cooperation projects in the areas of 3D modeling and visualization, geographic information systems, user interfaces, and web-based applications