通用率失真模型及码率控制算法

General rate distortion model and rate control scheme

为了提高码率控制的准确性,通过实验和理论推导,提出了一种通用率失真模型。在此基础上,定义了纹理复杂度,提出了纹理复杂度和图像块头用字节两自适应线性预测模型及宏块(组)比特分配模型;然后结合以上模型提出了一种新的码率控制算法。大量仿真实验结果表明,利用新提出的码率控制方法,测试序列的PSNR均有不同程度的提高,且码流曲线更平滑平稳,目标码率控制更准确。

To improve the accuracy of rate control, proposed a new general rate distortion model on the basis of theoretical analysis and extensive experiments. Meanwhile, this paper defined texture complexity, presented optimistic bit allocation model and two adaptive linear predictive models：texture complexity and overhead bit count predictive models. Based on above models, proposed a new rate control scheme for low bit-rate and real-time video coding applications. Simulation results show that the new rate control algorithm provides smoother bit stream, higher video quality and more accurate rate control than previous algorithms.