Embedded coding of images with regions of interest based on zerotree
Yunsong Li · Jounal of Xidian University · 2002
An embedded zerotree coding algorithm for the compression of the images with regions of interest is presented and a new weighted PSNR is also proposed to evaluate the performance of the reconstructed image. The scheme utilizes a simple and effective ROI mask generation using the zerotree structure without relying on the wavelet filters. The experimental results reveal that the PSNR results of the regions of interest are up to 4?dB higher than those by the oringinal embedded zertree algorithm, and the proposed weighted PSNR corresponds to the human visual impression.