Characterization of Breast Mass in Mammography using Image Texture Analysis

Abstract

Breast cancer is the most common type of cancer among women in the world. Mammography is regarded as an effective tool for early detection and diagnosis of breast cancer. In this study an approach is proposed to develop a computer-aided classification system to characterize breast mass from digital mammograms using IDL programming language by feature extraction for 9 features. The sample is 155 mammogram images and the data collected randomly from X-ray department at cancer diagnostic medical center. The study was conducted from April 2016 to March 2020. The proposed system consists of two steps. The first step is the feature extraction by using first order statistics using 3 features (mean-energystandard deviation)and the classification accuracy of breast tissues and tumors is for Tumor 96.8%, gland 57.9%, fat 98.9, While the connective tissue showed a classification accuracy 98.5%. The overall classification accuracy of breast area by using first order The second step is feature extraction by using higher order statistics (long run emphasis (LRE) , grey level non uniformity (GLN), run length non uniformity (RLN), Run percentage (RP), High Gray Level Run Emphasis (HGLRE) and Low Gray Level Run Emphasis (LRHGLE) ) and the classification accuracy of breast tissue and tumor showed a classification accuracy for tumor 88.9%, gland 98.9%, fat 86.3%, connective tissue 91.9%.The overall classification accuracy of breast area by using second order statistics 91.5%. Mammographic texture analysis is a reliable technique for differential diagnosis of breast tumors and breast tissue. Furthermore, the combination of imaging-based diagnosis and texture analysis can significantly improve diagnostic performance