1.Objective: Detect the area of asymmetric tissues on a pair of mammograms and then improve the accuracy of breast screening.
2.Key Modules: (1)’Image Calibration Module’ assists to improve the accuracy in lesion detection and recognition; (2)’Analytics Model Establishment Module’ determines whether a pair of mammograms is asymmetric or not, and which one is asymmetric to the other using deep learning; (3)’Symmetry Analysis Module’ detects the area of abnormal tissue which results in asymmetry in the mammogram.
3. Function: Our technique is capable of judging the asymmetry on a pair of mammograms, determining which breast is the asymmetric, and detecting the area of asymmetric tissues.