| Hi-index | 0.00 |
Breast cancer can be most effectively treated when detected during its early stages. Currently, the most effective method for early detection and screening of breast cancer is X-ray mammography. But the high number of false positives in mammography causes a lot of unnecessary biopsies. Sonography is an important adjunct to mammography in breast cancer detection and has been primarily useful for differentiating cysts from solid tumors. It has been shown that breast sonography is superior to mammography in its ability to detect focal abnormalities in the dense breasts of adolescent women. The accuracy rate of breast ultrasounds can reach a high level in the diagnosis of simple benign cysts and reduce the number of false positives. The lesions characterized as benign cysts do not require biopsies.This research focuses on developing a novel computer-aided diagnosis system of mass detection and classification in breast ultrasound images using statistical methods. First, we propose a novel, automatic enhancement approach to preprocess the regions of interest in breast ultrasound images. Next, a Markov random field segmentation algorithm is developed to segment the suspicious areas from the regions of interest. Third, three kinds of features, including textural features, fractal dimensions, and histogram-based features, are analyzed and extracted based on the suspicious areas and regions of interest, and a stepwise regression method is used to select an optimal subset of features. Fourth, based on the optimal subset of features, we develop a novel classification method based on the fuzzy support vector machines to classify the regions of interest as benign or malignant masses. Finally, two objective evaluation methods, five objective indices, and the receiver operating characteristic curve are used to evaluate the performance of the proposed computer-aided diagnosis system. The experimental results show that the proposed computer-aided diagnosis system greatly improves the five objective indices and the value of Az (the area under the receiver operating characteristic curve), as opposed to other classification methods and radiologists' assessments. Our method's ability to improve the accuracy of breast cancer diagnosis while reducing the number of unnecessary biopsies is a valuable tool for patients and radiologists alike.