A novel image retrieval algorithm based on color and distribution entropy of wavelet salient points
Zhiyong Zeng, Zhiyong An · Journal of Lanzhou University · 2006
Having analyzed the drawbacks of image retrieval based on interest points,a novel image re- trieval algorithm is proposed in this paper.using wavelet-based salient points.The algorithm extracts the salient points in wavelet domain and these salient points not only represent global variations as well as local variations;then the salient points are regarded as clues and color distribution entropy is designed,which takes not only the local color feature into consideration,but also the space distribution information of the salient points.The similarities are measured by the distance of color distribution entropy.With robustness to rotation and translation,the algorithm avoids the shortcoming of losing the location information in the traditional color histogram.Experimental results show that this algorithm is efficient for image retrieval.