项目名称: 基于超分辨率技术的视频重构与编码研究
项目编号: No.60803068
项目类型: 青年科学基金项目
立项/批准年度: 2009
项目学科: 医药、卫生
项目作者: 马思伟
作者单位: 北京大学
项目金额: 20万元
中文摘要: 为了应对超高分辨率视频应用所提出的高效压缩技术挑战,以及从低分辨低质量视频生成高分辨率高质量视频- - 以在高分辨率设备上得到更好的主观显示效果的技术难题,本项目提出进行基于超分辨率技术的视频重构与编码研究。在过去的三年里,通过对超高分辨率视频编码技术与理论进行深入研究,建立视频空时域相关模型,提出了高效的时/空间超分辨率算法及基于超分辨率技术的高效视频编码框架,其中提出的随机采样图像高质量恢复算法发表在VCIP2011上并获最佳论文奖。重要研究成果产出包括发表论文24篇,其中期刊论文4篇(本领域顶级国际期刊IEEE Trans. on Image Processing 2篇),SCI收录检索7篇,EI收录检索17篇,录用本领域顶级国际期刊IEEE Trans. on CSVT 3篇,1篇VCIP2011论文获最佳论文奖,申请专利5项。人才培养方面,已毕业博士生1名,硕士生2名。
中文关键词: 超分辨率技术;高分辨视频编码
英文摘要: With the fast development of hardware and multimedia communication technology, video applications become more and more widely, and the developing trend of video capture/display devices present two directions: the high resouliton, high quality devices for digital TV broadcasting, and the low resolution low quality devices due to the cost of hardware, power or space limitation, which can be used in high density surveillance, mobile handed communication devices. These applications ask for high effeiciency coding techniques for super high defintion video. Besides this, the super resolution techniques from low resolution low quality video to high resolution high quality video are also required. This project proposes to research super-resolution based video reconstruction and coding. Based on the spatial-temporal correlation model, a super-resolution based high efficiency coding framework is proposed.We have published 24 papers, including 2 IEEE T-IP papers. Three IEEE T-CSVT papers got to be adopted. One VCIP paper got the best paper award.
英文关键词: Super-resolution;high resolution coding;