Image retrieval based on texture features at sensitive area
Zhangwei Li · Journal of Zhejiang University of Science and Technology · 2010
The paper proposes a method of the color image retrieval based on the dynamic characteristics of the sensitive areas.In conjunction with the human visual system characteristics,the scheme uses k-means clustering in the wavelet transform domain to extract the image from the sensitive area.The dynamic texture features being obtained from sensitive area represent the image content,which are employed to calculate the similarities between images and to perform image categories and retrieval.Simulation results show that the images can be precisely found out by this method,thus the retrieval accuracy will be significantly improved.