Detection of Breast Cancer Using Digital Breast Tomosynthesis

Abstract

The tumor is an unwanted tissue that develops in breast. Breast cancer may include a swelling, change in shape of the breast, rashes on the skin, and the watery-like substance coming from the nipple—American Cancer Society (Breast cancer, 2011) [1]. As the cancer spreads, there may be different symptoms like pain in breast and swelling. In conventional two-dimensional mammography, overlapping of the tissues is very considerable problem. The new methodology digital breast tomosynthesis is used for the identification of the breast masses. The input images are preprocessed using adaptive median filtering technique which reduces the impulse noise. In the next stage, image is segmented using Gaussian mixture model (GMM) where GMM is a category of the clustering algorithm. Next, the images are subjected to feature extraction. And finally, it is considered for feature classification using probabilistic neural network (PNN) classifier.