泉州师范学院学报
泉州師範學院學報
천주사범학원학보
Journal of Quanzhou Normal College
2008年
4期
20~24
,共null页
颜色特征 纹理特征 特征融合 医学图像检索
顏色特徵 紋理特徵 特徵融閤 醫學圖像檢索
안색특정 문리특정 특정융합 의학도상검색
color feature; texture feature; feature fusion; medical image retrieval
为克服基于单一特征进行医学图像检索存在的局限性,文章提出了一种新的基于投影和纹理信息融合的检索算法,分析医学图像的特点,引入颜色投影并选择综合性能较高的特征进行融合,调整各特征的权重,最后运用加权欧氏距离进行相似性度量.实验结果表明,该算法检索速度快,具有较高的查准率.
為剋服基于單一特徵進行醫學圖像檢索存在的跼限性,文章提齣瞭一種新的基于投影和紋理信息融閤的檢索算法,分析醫學圖像的特點,引入顏色投影併選擇綜閤性能較高的特徵進行融閤,調整各特徵的權重,最後運用加權歐氏距離進行相似性度量.實驗結果錶明,該算法檢索速度快,具有較高的查準率.
위극복기우단일특정진행의학도상검색존재적국한성,문장제출료일충신적기우투영화문리신식융합적검색산법,분석의학도상적특점,인입안색투영병선택종합성능교고적특정진행융합,조정각특정적권중,최후운용가권구씨거리진행상사성도량.실험결과표명,해산법검색속도쾌,구유교고적사준솔.
Novel medical image retrieval algorithm based on the fusion of projection information and texture information is presented in this paper to overcome the limitations of a single feature in the description of medical image content. Through the analysis of the medical image feature, we study projection information, choose those signgle features which are suitable to the medical images and have high performance,adjust the feature weights,and use weighted Euclid distance to measure distance. Experimental results show that the search algorithm performs quickly and has a high precision.