Research and Development Department, Med Fanavarn Plus Co., Karaj, Iran
Breast cancer is the second leading cause of cancer death in women. Early detection and therapy are essential for the decline in breast cancer mortality rate. Mammography and ultrasound (US) are two conventional imaging modalities for breast cancer diagnoses. However, they might lead to unnecessary biopsy operations. The precise delineation of breast lesions’ borders is significant since determining the malignancy of a lesion is critically reliant on the lesion’s morphological features (e.g., shape, smoothness of boundary, etc.). Therefore, accurate detection of the lesion boundaries can assist in automated breast tumors’ classification. This study has utilized the Mask R-CNN model as a practical approach to segment breast lesions.