Statistical measures and criteria for ROI identification in breast mammograms

Breast cancer is one of the most common types of cancer in women worldwide. Studies proven that an early diagnosis of breast cancer can increase five year survival rate from 60% to 80+% [1]. That made screening programs a mandatory step for females. Therefore, radiologists have to examine a large number of images which may lead to missed breast lesions at early stage due to work load. Computer-Aided-Diagnosis (CAD) systems can be a powerful tool to overcome this problem by highlighting suspected lesions. However, this task is challenging also from CAD systems point of view due to difficulties in articulating and modeling patterns of abnormalities in a computational way. Many image processing methods were developed over the past two decades to help radiologists in diagnosing breast cancer. In this paper a new algorithm is introduced for Mammograms Region Of Interest (ROI) identification using statistical properties of mammograms. The proposed algorithm has been verified using 115 mammograms from the MIAS databases and other sources. Simulation results show that the proposed algorithm achieved 17% False Positive (FP) reduction on average vs best in class detection methods.