基于模糊支持向量机的面向语义图像检索算法*

为了缩减图像低层特征和高层语义之间的“语义鸿沟”，本文提出一种基于模糊支持向量机的面向语义图像检索(SBIR-FSVM)算法。在提取图像的低层特征的基础上，本文将最小隶属度模糊支持向量机引入到图像检索技术中，获取图像语义信息及消除传统支持向量机(SVM)在多类分类中产生的不可分区域，从而实现面向语义的图像检索。实验结果表明，本文提出的SBIR-FSVM算法与基于SVM的图像检索算法及综合多特征的基于内容的图像检索算法相比均有了显著的改进。

Semantic-based image retrieval algorithm using fuzzy support vector machine

To bridge the semantic gap between low-level features and high-level semantics, our paper introduced a new method called semantic-based image retrieval using fuzzy support vector machine (SBIR-FSVM): By extracting the low-level features of images and introducing the min-membership-function fuzzy support vector machine into image retrieval, the image semantic information was obtained and unclassifiable regions in traditional SVM in multi-class classification were avoided, thereby, semantic-based image retrieval was realized. Experiments show that the SBIR-FSVM algorithm our paper proposed is superior to the image retrieval algorithm based on SVM and the CBIR algorithm using multi-feature.