Hybrid Edge Based Deep Model for Accurate Mammogram Segmentation and Classification

Abstract

The mammogram images are used to identify breast cancer in the earlier stage. This will minimize the mortality rate in women. Now a days cancer is a rapid spreading disease, and it is hard to diagnosis accurately. It is hard for radiologist to find the correct region of spreading of disease. For this deep learning play a vital role in segmentation, extracting the feature and classify the image. In this paper four stages of findings were discussed to get the best result from the mammogram images. The pre - processing by Histogram Equalization is used to enhance the images of mammogram for better segmentation a & feature extraction. The second stage is to segment the images by ROI using Edge based segmentation. The third stage is to extract the feature using wavelet transformation to get the exact feature from the mammogram images. The fourth stage is to classify the image using Deep Resolute Neural Net. The accuracy of our proposed model’s performance was compared to that of currently used detection methods. The experimental result findings of our suggested method DR-NN model is superiority over the presently followed state-of – the art methods.