基于角点的动态加权多特征融合图像检索方法

Dynamic weighted multi-features fusion image retrieval method based on corner points

多特征融合的检索方法是图像检索系统中常用的方法。文章提出了基于角点的动态加权多特征融合的图像检索方法,首先,采用Harris算法获取图像的角点,由此计算图像的兴趣区域(图像的内容主体)。其次,在兴趣区域内提取图像的底层特征(颜色特征、纹理特征),对提取的特征进行归一化处理。最后,对多个特征分别与待检索图像进行相似度计算,利用动态加权进行相似度层的融合,将检索结果排序后提交给用户。在自建数据库上进行实验,表明该方法对提高图像的检索性能是有效的。

The retrieval method of multi feature fusion is a common method in the image retrieval system. In this paper, the image retrieval method based on the dynamic weighted multi feature fusion based on the corner point is proposed. First, Harris algorithm is used to obtain the corner of the image, and then the interest region of the image is calculated （the content body of the image）. Secondly, the underlying features （color features and texture features） of the image are extracted in the interest region, and the extracted features are extracted. The feature is normalized. Finally, the similarity of the multiple features and the image to be retrieved is calculated, and the similarity layer is fused by dynamic weighting. The retrieval results are sorted and submitted to the users. Experiments on a self built database show that the method is effective in improving image retrieval performance.