A New Method to Classify Breast Cancer Tumors and Their Fractionation

Autores

  • Omid Rahmani-Seryasat Department of Electrical and Computer Engineering, Hakim Sabzevari University, Sabzevar, Iran
  • Javad Haddadnia Associate Professor, Electrical and Computer Engineering Department, Hakim Sabzevari University, Sabzevar, Iran
  • Hossein Ghayoumi-Zadeh Department of Electrical and Computer Engineering, Hakim Sabzevari University, Sabzevar, Iran

DOI:

https://doi.org/10.5902/2179460X19428

Palavras-chave:

breast cancer, classification, growth areas, phase clustering

Resumo

In this paper, suspicious breast tumors were classified by using the neural network and the growth area method has been used for a fractionation of the benign or malignant areas of the normal tissue. Features extracted from input tissues are including statistical features and characteristics of spatial dependence. The advantage of this method is using of phase adaptive threshold based on entropy which leads to more accurate extraction of tumors and also corresponded with the nature of mammogram images. As a result, this method mimics of the human eye operation to detect abnormal masses. Database used in this paper is the MIAS mammogram database including 238 normal, benign and malignant mammograms. The accuracy obtained with 38 features is equal to 86.66% for detecting abnormal masses and 38.05 % for normal masses.

Downloads

Não há dados estatísticos.

Downloads

Publicado

2015-09-14

Como Citar

Rahmani-Seryasat, O., Haddadnia, J., & Ghayoumi-Zadeh, H. (2015). A New Method to Classify Breast Cancer Tumors and Their Fractionation. Ciência E Natura, 37, 51–57. https://doi.org/10.5902/2179460X19428