Feature weighting for cancer tumor detection in mammography images using gravitational search algorithm
Fatemeh Shirazi, Esmat Rashedi · 2016
Optimization methods have been widely used in image processing and computer vision. In this paper, k-nearest neighbor (KNN) and real-valued gravitational search algorithm (RGSA) are used to detect the breast cancer tumors in mammography images. GSA is used as a tool for optimization of the features weighting (FW) and tuning the classifier. FW-KNN based on GSA is employed to enhance the K-NN classification accuracy. The weighted features and the tuned K-NN classifier are utilized for detecting tumors. The obtained results show good efficiency of GSA-based FW-KNN classification for breast cancer tumor detection.