流形学习和相关反馈相融合的医学图像检索

图像检索是医学图像辅助诊断的基础，为了提高医学图像检索的正确率，提出一种流形学习和相关反馈相融合的医学图像检索算法(LLE-MF)。首先根据方块编码的思想提取颜色分量的信息熵，并利用邻域灰度共生矩阵提取纹理特征；然后采用非线性流形学习对颜色和纹理特征进行组合、降维处理，并采用欧式距离相似度量模型对图像初步进行检索，最后最小二乘支持向量机对初步检索结果进行相关反馈，并进行仿真测试。结果表明，相对于其它医学检索算法，LLE-MF不仅提高了医学图像的检索准确率，同时提高了医学图像的检索效率，可以准确地找到用户所需的图像.

Medicinal Image retrieval based on relevance feedback and manifold learning

age retrieval is the basis of medical image diagnosis, in order to improve the accuracy of medical image retrieval, this paper proposed a medical image retrieval based on manifold learning and relevance feedback retrieval algorithm (LLE-MF). information entropy of the color feature is extracted according to the block coding idea, and texture feature is extracted by neighborhood gray level co-occurrence matrix; and then the nonlinear manifold learning is used to select features from color and texture features, and Euclidean distance similarity is used to get the preliminary model of image retrieval, finally, he least square support vector machine is used to relevant feedback based on the preliminary search results, and the performance of the algorithm is tested by simulation test. The simulation results show that, compared with other medical retrieval algorithms, the proposed algorithm can extract the image feature extraction , not only improve the medical image retrieval accuracy, and improve the medical image retrieval efficiency, it can more accurately find the image required by the user.