| 一种残差-预测重构的视频分布式压缩感知实现方法研究 |
| |
| 引用本文： | 赵慧民,裴真真,才争野,王晨,戴青云,魏文国.一种残差-预测重构的视频分布式压缩感知实现方法研究[J].计算机科学,2017,44(6):317-321. |
| |
| 作者姓名： | 赵慧民 裴真真 才争野 王晨 戴青云 魏文国 |
| |
| 作者单位： | 广东技术师范学院计算机科学学院 广州510665 广州市数字内容处理及其安全性技术重点实验室 广州510665,广东技术师范学院计算机科学学院 广州510665 广州市数字内容处理及其安全性技术重点实验室 广州510665,广东技术师范学院计算机科学学院 广州510665 广州市数字内容处理及其安全性技术重点实验室 广州510665,广东技术师范学院计算机科学学院 广州510665 广州市数字内容处理及其安全性技术重点实验室 广州510665,广东技术师范学院计算机科学学院 广州510665 广州市数字内容处理及其安全性技术重点实验室 广州510665,广东技术师范学院计算机科学学院 广州510665 广州市数字内容处理及其安全性技术重点实验室 广州510665 |
| |
| 基金项目： | 本文受国家自然科学基金(61672008),广东省自然科学基金(2016A030311013,5A030313672),广东省应用型科技研发专项项目(2016B010127006,2015B010131017),广东省教育厅国际科技合作项目(2015KGJHZ021),广东省科技计划项目(2014A010103032)资助 |
| |
| 摘 要： | 为了降低计算成本并节约系统功耗,信号处理最新出现的理论-分布式压缩感知(Distributed Compressed Sensing,DCS)成为视频技术的应用焦点。为此,一种基于多假设预测的视频DCS(VDCS)方案被提出。在VDCS的解码端,当前帧的预测来自于以前重构的参考帧(CS帧),而残差作为重构条件用于改善视频的重构质量。实验结果表明,提出的残差-预测VDCS方法重构视频信号的峰值信噪比(PSNR)优于MH-BCS-SPL和传统的JSM-DCS处理方法。
|
| 关 键 词： | 分布式压缩感知 视频 残差 预测 重构 |
| 收稿时间： | 2016/12/21 0:00:00 |
| 修稿时间： | 2016/2/18 0:00:00 |
Video Distributed Compressive Sensing Research Based on Multihypothesis Predictions and Residual Reconstruction |
| |
| ZHAO Hui-min,PEI Zhen-zhen,CAI Zheng-ye,WANG Chen,DAI Qing-yun and WEI Wen-guo.Video Distributed Compressive Sensing Research Based on Multihypothesis Predictions and Residual Reconstruction[J].Computer Science,2017,44(6):317-321. |
| |
| Authors: | ZHAO Hui-min PEI Zhen-zhen CAI Zheng-ye WANG Chen DAI Qing-yun and WEI Wen-guo |
| |
| Affiliation: | School of Computer Science,Guangdong Polytechnic Normal University,Guangzhou 510665,China Guangzhou Key Laboratory of Digital Content Processing and Security Technology,Guangzhou 510665,China,School of Computer Science,Guangdong Polytechnic Normal University,Guangzhou 510665,China Guangzhou Key Laboratory of Digital Content Processing and Security Technology,Guangzhou 510665,China,School of Computer Science,Guangdong Polytechnic Normal University,Guangzhou 510665,China Guangzhou Key Laboratory of Digital Content Processing and Security Technology,Guangzhou 510665,China,School of Computer Science,Guangdong Polytechnic Normal University,Guangzhou 510665,China Guangzhou Key Laboratory of Digital Content Processing and Security Technology,Guangzhou 510665,China,School of Computer Science,Guangdong Polytechnic Normal University,Guangzhou 510665,China Guangzhou Key Laboratory of Digital Content Processing and Security Technology,Guangzhou 510665,China and School of Computer Science,Guangdong Polytechnic Normal University,Guangzhou 510665,China Guangzhou Key Laboratory of Digital Content Processing and Security Technology,Guangzhou 510665,China |
| |
| Abstract: | For a low-cost and low-power demand,distributed compressed sensing (DCS),an emerging framework for signal processing,can be used in video application,especially when available resource at the transmitter side are limited.Therefore,a novel video DCS(VDCS) scheme was proposed in this paper,where multihypothesis(MH) predictions of the current CS frame are generated from one or more previously reconstructed CS frame.Meanwhile at decoder side,the predictions are utilized as a residual signal to improve reconstructed video quality.Experimental results demonstrate that PSNR performances of the proposed VDCS scheme outperforms other methods,such as MH-BCS-SPL and traditional JSM-DCS. |
| |
| Keywords: | Distributed compressive sensing Video Residual Prediction Recovery |
|
|
点击此处可从《计算机科学》下载全文 |
|
