Content-Based Retrieval of Calcification Lesions In Mammography
In order to assist doctors to detect micro-calcifications,about the similar lesions retrieval problem of mammographicmicro-calcification cluster, we develop a new algorithm withmulti-feature fusion and relevance feedback based on the studyof single feature and feature fusion using single distance measureimage retrieving techniques, this method adopts multi-distancemeasure to calculate the similarity directing at different features.Experiment is based on mammography image database whichcontains 250 mammography images and each image containscalcification cluster, we verified the retrieval performance by theprecision -recall ratio (PVR) of feature fusion using singledistance, feature fusion using multi-distance and relevancefeedback. Experimental results show that the method has a betterretrieval result than these methods which based feature fusionwhich using single distance measurement.
Song Li-xin Chang Rui-feng Wang Qian Wang Li
Dept.of Electronics and Information Engineering Harbin University of Science And Technology Harbin , Dept.of gynecology Heilongjiang hospital Harbin , China
国际会议
成都
英文
1-4
2010-06-18（万方平台首次上网日期，不代表论文的发表时间）