Affiliation:
1. Govind Ballabh Pant Engineering College, India
2. Uttarakhand Technical University, India
Abstract
Breast cancer is one of the main health issues for women. This disease can be cured only if detected at early stages. Digital mammography is used to detect the malignant cells at an early stage. This article designs a methodology to detect the malignant tumors. The methodology is comprised of preprocessing feature extraction by Gabor and Law's feature extraction, and feature reduction by ant-lion optimization as well as a classification step using a SVM classifier which is implemented on the live dataset prepared through the Rajindra Hospital Patiala along with MIAS and DDSM datasets. The results of proposed techniques have been compared with three states of art techniques SVM based classification without feature reduction, PSOWNN i.e. PSO based reduction with a neural network as a classifier and binary gray wolf-based feature reduction with SVM classifier. The performance analysis proves the significance of the technique.
Subject
Computer Networks and Communications
Reference42 articles.
1. Variants of Particle Swarm Optimization in Enhancing Artificial Neural Networks.;J.Abdulkadir;Australian Journal of Basic and Applied Sciences,2013
2. MRI mammogram image classification using ID3 algorithm
3. A comparative study of fuzzy c-means algorithm and entropy-based fuzzy clustering algorithms.;S.Chattopadhyay;Computer Information,2012
4. Computer-aided characterization of mammographic masses: accuracy of mass segmentation and its effects on characterization
5. Application of wavelet based k-means algorithm in mammogram segmentation.;S.Dalmiya;International Journal of Computers and Applications,2012