Automated detection of spiculated masses using integrated method based on active contour

Abstract

Medical image processing techniques have been used for breast cancer diagnosis research in the last few years. The spiculated mass is a factors that indicates underlying malignancy. This proposes an automatic algorithm for speculated mass detection. The algorithm comprises efficient image processing steps. Removing the pectoral muscles and digital mammography background leaves only the fatty tissue and breast masses that are early priorities of this algorithm. Then automatic extraction of ROI is required. The proposed polynomial improves the quality of the ROI in term of intensity contrast. The initial models of active contour based on GGVF are constructed using Radon transform and the hierarchical clustering. The final shape of active model represents the irregular shape of spiculation. The numerical tests employing images from the digital database for screening mammography show good accuracy of our proposed algorithm for detecting spiculated masses.