Computer-Aided Masses Detection Algorithms Based On Digitized Mammograms

Breast cancer is a common malignant tumor. The key to lower risk of death from breast cancer is early diagnosis and treatment. Mammography is the first choice to diagnose breast cancer. However, even the experienced clinicians cannot easily discover all possible lesions in time. Therefore people have been proposed the application of computer technology to assist clinicians with breast cancer diagnosis to improve its accuracy. With the rapid development of computer technology, computer-aided masses detection based on digitized mammograms has been a research hotspot in early diagnosis of breast cancer.Masses are the major indications of breast cancer in mammograms. In this paper, a computer-aided detection algorithm is proposed to automatically detect masses in mammograms. Firstly, breast region extracting algorithm and peripheral breast gray value correcting algorithm are proposed in the mammogram preprocessing step. These two algorithms aim at obtaining the breast region in mammograms and enhancing the dark portions due to the reduction of thickness along the peripheral breast. According to the characteristics of multi-layer annular filtering for breast masses, a fast algorithm for searching ROI (regions of interest) is proposed, which can fast search suspected massive regions and reduce a large number of false-positive regions by using partial features of masses. Then combining automated seeded mass segmentation algorithm, which can automatically segment suspected massive regions when the center of ROI as seed points. Finally, to minimize its false-positive regions, some features are extracted from the segmented massive regions to classify these regions.The experimental results demonstrate that the proposed algorithm works well, which can be applied to effectively detect massive regions in mammograms.