无反馈分布式视频编码中Wyner-Ziv帧码率控制算法

基于分布式视频编码系统,提出一种Wyner-Ziv帧的码率控制算法.首先,通过理论推导和实验,提出了针对Wyner-Ziv帧不同系数带的率失真估计模型,并通过训练学习的方法针对不同系数带分别确定模型参数;然后,使用提出的率失真模型估计当前帧的编码码率,并结合目标码率约束确定最优量化矩阵;最后,为每个比特面分配相应数目的校验位,完成LDPC编码并输出.实验结果表明,在相同目标码率下,与现有无反馈码率控制算法相比,使用本算法可提高解码恢复视频峰值信噪比0.3～1dB,且编码端运算复杂度增加很小.

Rate control algorithm for the Wyner-Ziv frame for no-feedback distributed video coding

A rate control algorithm for Wyner-Ziv frames for no-feedback distributed video coding system is proposed.Based on the rate-distortion theory and experimental results,the algorithm first establishes the rate-distortion estimation model(RDEM) for each coefficient band of Wyner-Ziv frames,and determines the model parameters by training.With the proposed RDEM,the algorithm is able to estimate the coding rate for each coefficient band of a Wyner-Ziv frame and choose the optimal quantization matrix accordingly to meet the target rate constraint.Finally,the parity rate estimation for each bit plane of each coefficient band is performed to accomplish the LDPC coding.Compared with the existing no-feedback rate control algorithm,numerical results show that the proposed method can increase the average PSNR(peak signal-to-noise ratio) of reconstructed images by 0.3 to 1 dB with little complexity incurred at the encoder.