Date of Publication :8th September 2017

Abstract: Earlier is the diagnosis of a disease, better is the rate of recovery. So far as the fatal disease like breast cancer is concerned, it’s early diagnosis may lead to improve the rate of care and thereby survival of a patient. Generally, breast cancer detection and analysis starts from capturing the Mammogram of the effected breast region. In this paper, an automated diagnosis scheme has been proposed for detecting the presence/ absence of breast cancer from such mammograms. Suitable pre-processing is applied to input mammogram images. For the feature extraction, the gray level co-occurrence matrix is framed out of the preprocessed image. The AdaBoost technique has been used for the purpose of feature selection. Classification is carried out with the help of the state-of-the-art Random-forest classifier. For the purpose of validation, the mammography image analysis society (MIAS) database has been taken into consideration. Satisfactory classification rate of 94% is achieved through the proposed scheme.

Reference :