Detection and classification of clusters of microcalcifications on mammographic images

An algorithm for detecting and classifying clusters of microcalcifications on mammograms is proposed. A feature of the algorithm proposed here is its potential for application to different types of calcifications and clusters of calcifications (both benign and suspicious). At the same time, vascular calcifications, which often give false positive results in algorithms, are analyzed separately, and a solution to this problem is proposed. The effectiveness of the proposed methods was assessed using a database of mammograms from patients with verified diagnoses. The classification algorithm achieved 96.15% accuracy, 95.32% specificity, and 98.21% sensitivity.