A Method of Region Feature Extraction and Representation in Content-based Image Retrieval

摘要

A novel approach is proposed for region multifeature extraction, representation and similarity matching in content-based image retrieval. The effectiveness of image segmentation is key job to extract and represent the region features. The main benefit of this algorithm is that the proposed method of image segmentation adopts an effective combination of global and local thresholding. The experimental results shows that the proposed method is feasible and effective.