Using Shape Circularity for Breast Cancer Detection

摘要

Background/Objectives: This study aims to automate breast cancer detection to either benign or malignant. This process helps physicians for early detection of the tumor. Methods Analysis: We propose a novel method for breast cancer detection called CARA. The proposed method is based on Circular ROI (Region of interest) of the tumor. The circularity of the tumor determines the tumor type either benign or malignant. Findings: The proposed CARA method achieves high accuracy rates compared to other previous work. CARA exceeds other previous work by more than 20% in accuracy. Moreover, CARA outperforms other previous work in all performed trials. Additionally, the proposed CARA method outperforms other previous work with respect to execution time. Improvements/Applications: Automatic classification of breast cancer to either benign or malignant in order to help physicians for early detection of the tumor.