Content-based image retrieval technology using multi-feature fusion

Abstract Due to the diversity of the image content, different images have different focuses, image retrieval system based on single feature has a lower performance, and it cannot apply to all images, so an image retrieval method using multi-feature fusion is proposed. In this method, the color moment in RGB color space in combination with the color histogram in HSV color space is used for color feature extraction, the improved Zernike moments are used for shape feature extraction, and the gray level co-occurrence matrix is used for texture feature extraction, then combining these three features. Finally, respectively using color features, shape features, texture features as well as the fused features for image retrieval, the experimental results show that the image retrieval method based on multi-feature fusion has better retrieval performance.