A New Classification Method based on Calcification Shapes in Mammographic Images

• Main Authors: Yu Hsin Hsien, 余信賢
• Other Authors: P.W. Huang
• Format: Others
• Language: zh-TW
• Published: 2005
• Online Access: http://ndltd.ncl.edu.tw/handle/89145759663925228521

碩士 === 國立中興大學 === 資訊科學研究所 === 93 === When most mammograms have been digitized, how to effectively assist the doctors to diagnose a large number of materials rapidly has become an important task. Due to the minimal area of calcification of mammogram, the shape of calcification is an important basis to diagnose that the calcification is benign or malignant. It will be very meaningful for doctor if we can classify the shape of calcification after we segment the calcification of mammogram. It makes the doctors get one more reliable data to diagnose the pathology of mammogram. This research is aimed at each calcification of mammogram image to determine and classify it. Because the area of calcification is usually minimal, general techniques of determining and classifying the shape are not easy to work out. Therefore, I propose a way called decision tree and appropriately constitute some simple features of shape through it. Then I discriminate the calcification of different shapes by Bayes classifier. The result of the experiment has proved that such tactic can reach good effect of classifying.