Image Retrieval Based On Color And Texture

With the continuous development of information technology, images as an important carrier of information are growing rapidly. How to efficiently retrieve image from huge resources is becoming a hot topic. Content-based image retrieval technology solves the limitations of traditional text-based retrieval techniques and improves the performance of retrieval greatly. Because single feature of the image has different shortcomings, the combination of more than one feature together for image retrieval is becoming more and more important.The key technologies of feature extraction and fusion technology are studied. The main contents include:1. A new method based on block truncation coding and gray level co-occurrence matrix was proposed in the paper. Based on the idea of block coding, the information entropy of each color component is extracted as color feature. Experimental results show that the proposed algorithm achieved better retrieval results than the other operators.2. Based on the combination of traditional GLCM and the statistical moments of the neighborhood, the neighbor statistic moment level co-occurrence matrix was introduced as texture feature. Furthermore, it was fused with color feature based on weighted method for multi-feature image retrieval. Experimental results denote that the proposed method achieved better retrieval results.3. The genetic algorithm was adopted to set the best weight values by the optimization of image feature in the paper. The experimental results show that the method has strong learning ability and the feature weight can be assigned automatically. On the other hand, the new method can improve the retrieval performance of image retrieval.