Design of image information retrieval system
Wei Fa-qiang · Jisuanji gongcheng yu sheji · 2007
Adopting methods that based on the images frequency characteristics matching and pattern recognition to retrieve the relevant information in the images is investigated.First,wavelet analysis is used to extract the image characteristic.Because wavelet analysis has multi-resolution analysis advantages,and it can extract frequency characteristic while keeping space information of the image,so it is used to extract the characteristic of images.Then accord to the computing complexity,the advanced two-dimension correlation algorithm and the two-dimension dynamic programming elastic algorithm are used to match and recognize the images.The system is implemented by using hybrid programming with VC++ and Matlab,and testing experiment is performed.The results of experiment show that the system has good query performance.