摘要
结合人眼视觉感知特性,提出一种面向对象基嵌入式编码的码率控制算法。首先结合人眼视觉感知特性估计各对象的重要性并确定其编码优先级;然后依据优先级对视觉对象进行位平面建模和熵编码,得到各自独立的码流;最后基于率失真优化理论在给定码率下对对象码流进行优化截取和重装。仿真实验结果表明,所提算法可以对不同重要对象进行差异化编码和传输,与基于PCRD算法的码率控制方法相比,新算法能提高重构图像的整体视觉效果。
A new rate control algorithm for object-based embedded coding was proposed by incorporating the character-istics of human visual systems（HVS）.Firstly,the importance and coding priority of each visual object were estimated.Then,bit-plane modeling and entropy coding were implemented for each object based on the coding priority and its cor-responding bit stream was outputted.Finally,bit streams of visual objects were truncated and reassembled based on the rate-distortion optimization principle under the given bit rate.Experimental results reveal that the proposed algorithm can encode and transmit different important objects with different strategies.Compared with the PCRD algorithm,the pro-posed algorithm can improve the overall visual quality of the reconstructed image.
出处
《通信学报》
EI
CSCD
北大核心
2012年第4期115-120,共6页
Journal on Communications
基金
国家自然科学基金资助项目(60902066
60872094
60832003)
浙江省自然科学基金资助项目(Y107740)
浙江省教育厅科研基金资助项目(Y201016875)~~
