Content-Based Scalable Multi-View Video Coding Using 4D Wavelet

Abstract

Most existing scalable multi-view video coding (SMVC) can provide temporal, spatial, SNR as well as view scalabilities, but not support content scalability. In this paper, a region of interest (ROI) coding is developed for scalable multi-view video codec to show the high scaling capability in terms of content scalability. The key component of this algorithm is that the bit-planes of the ROI can be lifted according to users’ requirements. And in this way the ROI can be allocated more bits even at low bit-rate and can secure more bits to get a better visual quality. As a result of our experiments, the peak signal to noise ratio (PSNR) within ROI can be improved by 2~4dB and the subjective visual quality of ROI in the proposed scheme can be improved compared to the conventional SMVC algorithm.