Tsallis entropy extraction for mammographic region classification

Abstract

Breast cancer is the second disease responsible for women’s death in the world. To reduce the number of cases, screening mammography is used to detect this disease. To improve exam accuracy results, computer-aided systems (CAD) have been developed to analyze the mammography and provide statistics based on image features extracted. This paper presents a novel approach for a computer-aided detection system (CADe) based on Tsallis entropy extraction from quantized gray level co-occurrence matrix (GLCM) from mass images. A comparison study is presented based on a feature extraction scheme using weigthed Haralick features. The best result accuracy rate was 91.3% from Tsallis entropy based on GLCM matrix using 24 feature measures.