基于多参考帧假设优化的压缩感知重构算法  

Reconstruction algorithm based on multi-reference frames hypothesis optimization for compressive sensing

作　　者：阔永红[1] 王薷泉 陈健[1] 

机构地区：[1]西安电子科技大学通信工程学院,陕西西安710071

出　　处：《通信学报》2017年第12期1-9,共9页Journal on Communications

基　　金：国家自然科学基金资助项目(No.61771366);"111"计划基金资助项目(No.B08038)~~

摘　　要：在多假设分布式压缩视频感知系统中,多假设的质量对重构性能意义重大。现有工作中,对于多假设集合获取的研究并未得到关注。提出一种多参考帧假设集合优化选择(MRHO)算法,增加参考帧数目以扩大假设选择范围,通过假设优化选择,在相同假设集合尺寸下提高了集合质量。仿真表明,MRHO算法有效提高了视频重构质量。In multi-hypothesis based distributed compressed video sensing systems, the quality of the multi-hypothesis set has important influence on the reconstruction performance of decoder. However, the acquiring of the hypothesis set has not been concerned in existing works. A reconstruction algorithm based on multi-reference frames hypothesis optimiza- tion （MRHO） was proposed. This algorithm expanded the selection of hypothesis vectors by increasing the number of reference frames. The quality of the prediction set was improved by hypotheses optimization selection under the same size with the original hypothesis set. Simulation results show that the proposed MRHO algorithm effectively improves the reconstructed quality of the distributed compressed video sensing scheme.

关 键 词：压缩感知 分布式压缩视频感知 多假设集合优选 多参考帧选择 

分 类 号：TP919.8[自动化与计算机技术]