Investigation on performance analysis of support vector machine for classification of abnormal regions in medical image

Abstract

The most common malignancy in Indian women is breast cancer. However, cancer can be detected earlier with mammography. Computer assisted diagnostic (CAD) techniques are a boon to the medical industry, and these techniques are designed to help physicians make a diagnosis. It presents a new CAD system for the detection and classification of mammographic abnormalities. The proposed work is divided into four main stages: pre-processing, segmentation, feature extraction, and classification. The pre-treatment phase aims to eliminate unwanted noise and make the mammogram suitable for the next process. The purpose of the segmentation phase is to highlight areas of interest for the continuation of the process. Extraction is the main step in which you need to extract texture elements from the region of interest. In this work, pseudo-grain moments are used to extract features due to noise tolerance and descriptive ability. Finally, a support vector machine is used as a classifier to distinguish between malignant and normal mammograms. The performance of the proposed work is carried out by different experiments and the results are satisfactory in terms of accuracy, specificity, and sensitivity.