基于IAFSA优化权值的医学图像检索  

Medicinal image retrieval based on improved artificial fish swarm algorithm optimizing weights

作　　者：石晓艳[1] 刘淮霞[1] 于水娟[1] 

机构地区：[1]安徽理工大学电气与信息工程学院自动化系,安徽淮南232001

出　　处：《计算机工程与设计》2014年第11期3961-3966,共6页Computer Engineering and Design

基　　金：安徽省教育厅自然科学研究基金项目(KJ2013B206);安徽省教育厅自然科学研究基金项目(KJ2013B207)

摘　　要：为提高医学图像的检索结果,提出一种基于改进人工鱼群算法的医学图像相关反馈检索方法 (ISAFA)。提取医学图像的颜色、纹理、形状特征,采用相似度量模型得到图像初步检索结果,根据用户的反馈信息,采用改进人工鱼群算法对特征权值进行调整,以获得更加理想的检索结果。检索结果表明,ISAFA提高了医学图像的检索准确率和检索效率,可以找到满足用户需要的医学图像。To improve the retrieval accuracy of medical images, a medical image retrieval method based on the improved artificial fish swarm algorithm optimizing the weights of features was proposed. Firstly, the color, the texture and the shape of medical images were extracted, and then the similarity measurement model was used to obtain the preliminary retrieval results. Finally, the improved artificial fish swarm algorithm was used to adjust the weights of features according to the feedback information of users to improve the image retrieval accuracy. The results show that, compared with other medical retrieval methods, the proposed method can improve the retrieval accuracy and efficiency of medical images, so it can accurately find the medical image required by the user.

关 键 词：改进人工鱼群算法 医学图像 相似度度量 权值调整 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]