面向分级B帧编码的分级量化技术

Hierarchical quantization technique based on hierarchical B pictures coding

提出了一种面向分级B帧编码的分级量化技术,通过发掘同一个Gop中B帧之间的时域相关性,按金字塔顺序在级别不同的B帧之间相应分配权重不同的量化步长,达到编码优化的目的。仿真实验证明,所提出的分级量化技术比传统的使用同一个量化步长因子的量化机制在保证相同重构图像质量的前提下,能较大幅度的节省码流。

This paper proposes a hierarchical quantization technique based on hierarchical B pictures coding.By efficiently exploiting the motion information within the B pictures in the same Gop and allocating different quantization steps between B pictures of different ranks according to hierarchical order,this paper achieves the goal of coding optimization.Simulated results show that the proposed technique saves the coding bitrates better than the traditional technique in the same quality of reconstructed pictures.