Hybrid Segmentation and Feature Extraction Approach to Detect Tumour Based on Fuzzy Rough-in Mammogram Images
M. Poomani Alias Punitha, K. Perumal · Procedia Computer Science · 2019
Tumor classification plays a significant area of research in mammogram images. In this paper, we introduce the tumor classification method in mammogram images by using Fuzzy rough set theory (FRST) and it offers an accurate approach of texture and feature extraction. The core purpose of deploying FRST is feature extraction which is achieved by using a quick reduct algorithm which helps to identify the tumor without loss of pixels in a short period. Fuzzy rough instance selection (FRIS) is applied to remove the noise from the mammogram image and finally the combination of fuzzy-rough nearest neighbor (FRNN) method is used in segmentation. The results obtained using the proposed methods are compared in various performance measures such as accuracy, sensitivity and specificity are calculated accurately.