An efficient content-based rate control for 3-D wavelet-based scalable video coding

Abstract

In this paper, an efficient content-based rate control scheme is proposed in the context of rate-distortion optimization for 3-D wavelet-based scalable video coding (SVC). Despite the prevalence of the conventional rate control methods in DCT-based SVC, the underlying formulation is not directly applicable to 3-D wavelet-based SVC. Performing optimal rate control with 3-D wavelet coding methods is difficult because energy is not conserved after applying the motion-compensated temporal filtering (MCTF) technique. To address the inherent challenges, a thorough analysis of the distortion relationship between temporal subbands and reconstructed frames is explored for controlling the video quality fluctuation. Although the quantization distortion of the reconstructed frame is assessed by tracking the filter-based quantization error along the temporal decomposition trees, it fails to consider the pixel connectivity status and the temporal subband content characteristic that ultimately degrades the video coding performance. Based on our statistical investigation, an adaptive MCTF decomposition level selection is first developed to enable more accurate motion compensation than the original MCTF. Then, the mutual information, the gradient per pixel, and the texture homogeneity are effectively utilized to measure the subband content and enhance the rate control performance. Experimental results show that the proposed scheme can reduce the bit estimation error and improve the video quality with smoother PSNR variation as compared to three other benchmark rate control schemes.