Computer-Aided Diagnosis of Breast Lesions

摘要

The long-term goal of our research is to develop computer-aided diagnosis (CAD) techniques to improve the detection and diagnosis of breast cancer. The hypothesis to be tested in the present project is that radiologists' ability to differentiate malignant from benign breast lesions can be improved by integrating radiologists' perceptual expertise in the interpretation of mammograms with the advantages of automated computer classification. This project has 3 objectives: 1. To combine radiologist-extracted Breast Imaging Reporting and Data System (BI-RADS) features with image features extracted by a computer to classify malignant and benign clustered microcalcifications in mammograms. 2. To optimally combine radiologists' diagnosis with the result of computer classification. 3. To optimize computer classification for full-field digital mammograms.