多特征融合及SVM相关反馈技术在教育资源图像检索中的应用

首先介绍了教育资源图像检索系统的结构模型和基本模块,然后介绍了图像特征提取技术,在颜色特征提取中主要应用了分块直方图法,在纹理特征提取中应用了Gabor小波变换法,在形状特征提取中主要应用了不变矩法,通过3种特征的权重分配将多特征融合进行图像初次检索,然后采用基于SVM的相关反馈技术进行多次检索。实验证明,多特征融合和相关反馈技术的引入有效提高了教育资源中图像检索的查准率。

Application of multi-feature integration and relevance feedback based SVM in image retrieval of educational resources

The paper describes the image retrieval system of educational resources and basic module structure model firstly, then introduced image feature extraction techniques. In color feature extraction the authors mainly used sub-block histogram method, and the Gabor wavelet transform method was used in texture feature extraction, while the invariant moment method was used in shape feature extraction. Multi-feature was integrated in the first time image retrieval through weight distribution of three characteristics, and then SVM-based relevance feedback technique was used in multiple image retrieval. The experiments indicate that this method can improve the image retrieval precision rate of educational resources effectively.