文档介绍：Content-Based Retrieval of Calcification Lesions In
Mammography
Song Li-xin, Chang Rui-feng, Wang Qian Wang Li
Dept. of Electronics and Information Engineering Dept. of gynecology
Harbin University of Science And Technology Heilongjiang hospital
Harbin , China Harbin , China
Lixins99@ Wangli1439@

Abstract—In order to assist doctors to detect micro- calcifications, query point movement, for improving system performance.
about the similar lesions retrieval problem of mammographic Although CBIR has been many applications proposed for
micro-calcification cluster, we develop a new algorithm with several applications abroad, one encounters a “semantic gap”
multi-feature fusion and relevance feedback based on the study between the quantitative features used to represent the images
of single feature and feature fusion using single distance measure and the interpretation of the images by users who are experts in
image retrieving techniques, this method adopts multi-distance the domain of application [5]. This leads to the need to guide
measure to calculate the similarity directing at different features. the retrieval algorithm by incorporating the user’s judgment of
Experiment is based on mammography image database which similarity and the relevance of each retrieved [6, 7].
contains 250 mammography images and each image contains
calcification c