Classification of salient dense regions in mammograms based on the minimum nesting depth approach

In this study, a novel method for classifying salient dense regions in mammograms is proposed. The method respectively includes detecting threshold based local maximum regions, eliminated with the decision tree process , computing features and minimum nesting depths for candidates of region of interests and finally classification by using Support Vector Machines ( SVM ). Experimental results demonstrate that the proposed method achieve good performance for detecting masses in mammogram.