Image Retrieval Method Based On Fuzzy Correlation Of Color Histogram And NSCT

With the development of multimedia technology, the number of images shows a exponentially growing trend. An image is a digital information carrier which has rich content, intuitive performance, no language restrictions and convenient exchange. The image retrieval technology is used to retrieve desired image from the mass image data quickly and efficiently.Based on the deep research and study of color and texture features in the content-based image retrieval technology, a image retrieval method based on fuzzy correlation of color histogram and NSCT is proposed, and multiple sets of compared experiments are designed to confirm their superiority. The main contents are as follows:Firstly, the paper focuses on key technologies of the system, such as feature extraction, similarity measure and evaluation criteria and so on.Secondly, the fuzzy set theory can well-described image’s uncertainties, and this paper uses the fuzzy set theory as the similarity measure method. Similarity is calculated by fuzzy membership function of the fuzzy set theory, and it is determined by a level fuzzy relationship, and the experiment proves that this method used to measure the similarity has very good results.Again, the existing image retrieval based on color feature is improved. The color histogram is used to extract color features of the image, then the fuzzy set theory is used to measure similarity, eventually retrieval results are returned. Compared with the image retrieval experiment based on cumulative color histogram’s Euclidean distance-related, it can be seen from the experimental results that the image retrieval method based on fuzzy correlation of color histogram has higher retrieval accuracy.Then, the existing image retrieval based on texture features is improved. NSCT is used to extract texture features of the image, then the fuzzy set theory is used to measure similarity, eventually retrieval results are returned. Compared with the image retrieval experiment based on contourlet transform’s Euclidean distance-related, it can be seen from the experimental results that the image retrieval method based on fuzzy correlation of NSCT has higher retrieval accuracy.Finally, these two methods are combined, then the method of the comprehensive features is used to retrieve image. Approach of comprehensive feature whose value setting will affect the image retrieval effect, making the color and texture features can complement and improve the accuracy of image retrieval. Experimental results show that this integrated features approach has better retrieval precision than the retrieval method of single feature, and than any other integrated features method.