Research on Image Retrieval Based on Local Features

Abstract

In the paper, we research on content-based image retrieval based on local image features. Local image features can extract useful information while avoid image segment. Firstly, we use some effective and classic texture features based on wavelet transform's subbands which coincide with human eyes perception. Secondly, we analysis local features for image retrieval. In order to extract the salient points exactly, we used a famous improved salient point detector. The experimental results demonstrate that the method is effective.