Research On And Implementation Of Content-Based Image Retrieval Technologies

As an important part of multimedia information retrieval. Content-Based Image Retrieval(CBIR) is being widely used in every field of Internet. More and more researchers achieved many breakthroughs and developed many business&research systems. For further study, this paper presented a new image retrieval method Dominant Color Features and Color Ring according to the one-to-one relation of image and its moments. With regard to texture, the paper used wavelet decomposition theory and advanced a new texture-extraction algorithm Wavelet Energy Covariance Matrix. In addition, statistical experiment accomplished in the paper proved that texture image can be decided according to its standard variance. At the last part of theory study, System performance criterions are deeply studied. Based on these theories and algorithms, CBIR System design and implementation are paid more attention. It is believed that design model and implementation scheme discussed in the paper would provide great reference &practice value for future CBIR application.