画像解析による超音波乳腺腫瘤像の計量診断 : 特徴パラメータの検討

概要

論文の詳細を見る
The purpose of the research is in the development of quantitative diagnosis system of a ultrasonic breast tumors image by computer image analysis. This report discusses a computer aided diagnosis system for the differential diagnosis of breast cancer, fibroadenoma, and cyst. The 20 quantitative feature parameters are calculated by image processing for a B-mode ultrasonic image of breast tumor. First, the parameters were examined quantitatively by using student t-test. Next, the combination of five was made from the three kind of tumor and a parameter effective each doing distinction was selected by stepwise discriminant analysis. And, linear discriminant coefficient of the selected parameter was calculated. The internal data breast cancer 50 examples, fibroadenoma 5l examples, and cyst 49 examples were distinguished by using the linear discriminant. As a result, the diagnosis accuracy between the breast cancer and the fibroadenoma, the breast cancer and the cyst was 100%, 100% though the diagnostic accuracy between the fibroadenoma and the cyst was 82.4%, 87.8%, low correct answer rates. It is believed that the automated diagnosis system can be constructed with the above-mentioned examination.
鈴鹿医療科学大学の論文
1994-05-01