Abstract: Breast cancer is a high incidence of malignant diseases in women and mammography is the main basis for diagnosis. At present, because of the tumor edge feature extraction of micro tumor detection methods in breast X-ray images is not effectively, the false positive results overhigh, which affects the diagnosis and treatment of patients. Therefore, the micro tumor detection method in breast X-ray images based on edge geometric features was studied. According to the morphology principle, the breast X-ray image was enhanced. HOG feature extraction method was used to obtain the edge geometric features of breast micro tumors, and constructed a sparse representation Dictionary of tumor edge geometric features. Using edge dictionary and subtraction technology to complete microtumor detection. The experimental results showed that compared with the traditional detection methods, this method could effectively reduce the number of false positives and improve the detection accuracy of micro tumors. Therefore, the method in this paper has high feasibility.

Key words: breast X-ray images, micro tumor detection, HOG feature extraction method, edge geometry