With the development of multimedia and internet technology, the application of the image is more and more extensive. How to retrieve necessary information from large amount image information efficiently and quickly needs to be solved urgently. Traditional image retrieval methods are no more appropriated for large image databases. In order to overcome the limitation of the traditional searching method, the CBIR has become one of the hot research areas in image domain.As for this study area, in this paper, we mainly focus on research and study the technique of CBIR based on color,texture and shape feature, and we manage to conclude and summarize algorithms for each feature extraction and representation. Firstly, enter on the color space selection and color feature extraction, we describe five different color spaces, discuss the color quantization,histogram definition and the application of color histogram in image retrieval. We have detailed description on texture definition and texture characters, introduced several texture analysis methods, including co-occurrence matrix,Gabor-wavelet feature, the experimental results show that Gabor-wavelet feature has good result on the pictures which background is simplex. We also describe Fourier shape descriptors and moment invariants in shape feature.Finally, we construct a content-based image retrieval system innovatively which composing HSV color histogram feature,co-occurrence matrix texture feature and moment invariants. We consider the complementarities of each feature fully which can learn from others' strong points to offset one's weakness in the methods of features extraction, and design an intelligent image retrieval platform which can compose each feature to retrieval. The experimental results show that the combing feature retrieval has the batter retrieval performance than the single feature retrieval. |