基于投影和纹理信息融合的医学图像检索

Medical Image Retrieval Algorithm Based on the Fusion of Projection Information and Texture Information

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.

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.