[1]周 亮,周 芳.基于Bayesian理论的压缩视频超分辨率重构[J].南京师范大学学报(工程技术版),2005,05(03):036-40.
　ZHOU Liang,ZHOU Fang.Compressed Video Super-Resolution Reconstruction Based on Bayesian Theory[J].Journal of Nanjing Normal University(Engineering and Technology),2005,05(03):036-40.
点击复制

基于Bayesian理论的压缩视频超分辨率重构
分享到：

南京师范大学学报（工程技术版）[ISSN:1006-6977/CN:61-1281/TN]

卷:
05卷
期数:
2005年03期
页码:
036-40
栏目:
出版日期:
2005-09-30

文章信息/Info

Title:
Compressed Video Super-Resolution Reconstruction Based on Bayesian Theory
作者:
周  亮1;  周  芳2
1. 南京邮电大学信息产业部与江苏省图像通信与图像处理重点实验室, 江苏南京210003;
2. 安徽工业大学电气信息学院, 安徽马鞍山243002
Author(s):
ZHOU Liang1;  ZHOU Fang2
1.Ministry Key Lab of IC and IP, Nanjing University of Posts and Communications, Jiangsu Nanjing 210003, China; 2. School of Electronic and Information, Anhui University of Technology, Anhui Maanshan 243002, China
关键词:
超分辨率;  压缩视频;  图像重构;  贝叶斯估计
Keywords:
super-reso lution;  com pressed v ideo;  im age reconstruction;  Bayesian estim ation
分类号:
TP391.41;
摘要:
针对传统算法不能很好解决压缩视频的超分辨率图像重构问题,提出了一种基于Bayesian估计的压缩视频超分辨率重构算法.首先建立起一个从原始图像到压缩视频的成像模型,然后在此模型基础上运用Bayesian估计理论,在最大后验概率准则下表述该问题.通过综合使用CCD(CyclicCoordinateDecent)和SA(SuccessiveApproximations)等方法,从理论上给出了压缩视频超分辨率重构问题的一般解决方法.实验结果表明,此方法不仅在峰值信噪比(PSNR)和重构效果对压缩视频有较大提高和明显改善,而且算法易于扩展,具有广泛的应用范围.
Abstract:
In th is paper, a new a lgo rithm of compressed v ideo super- reso lu tion resto ra tion based on Bayesian esti m ation is propo sed to reso lve the problem o f the traditiona l resto ra tion me thods wh ich cannot be we ll applied to the com pressed v ideo. A t first, an acqu isition o f the com pressed v ideo model is proposed, then the super-reso lution prob lem is form ulated w ith in the Bayes ian fram ew ork and theMAP criter ion, and fina lly, a universa l so lu tion o f the prob lem is presented by integrating the CCD w ith SA. The results of the experim ent dem onstrate that the a lgo rithm not on ly perform s better than the trad itiona l ones in term s of PSNR and restoration effect, but also has the charac teristics o f easy ex tension.

参考文献/References:

[ 1] Cheesem an P, kanefsky B, H ason R. Super-resolved sur face reconstruction from m ultip le im ag es[ J]. Tech Rep NASA, 1993, 21( 1): 269 -276.
[ 2] Schu ltz R R, Stevenson R L. Extraction of high reso lution frames from v ideo sequence [ J]. IEEE Trans on Im age Pro cessing, 1996, 9( 5): 996 -1011.
[ 3] H unt B R. Super-Reso lution o f im ag es: A lgorithm, Pr inci p les, Perform ance [ J]. In ternational Journa l of Imag ing System s and Techno logy, 1995, 14( 6): 297- 304.
[ 4] E ladM, FeuerA. Restoration o f a sing le superreso lution im age from several b lurred, no isy and undersamp led m eas- ured im ages[ J]. IEEE Trans on Im age Processing, 1997, 6( 12): 1646- 1658.
[ 5] E ladM, FeuerA. Superreso lution restoration if an im age sequence: adaptive filter appro ach [ J]. IEEE T rans on Im age Pro cessing, 1999, 12( 3): 387 -395.
[ 6] Patti A J, A ltunbasak Y. Super-reso lution im ag e estim a tion for transform coded v ideo w ith app lication to m peg [ J]. Pro c IEEE Int Conf Im age Pro cessing, 1999, 5( 3): 179 -183.
[ 7] Chen D, Schu ltz R R. Ex traction o f high-reso lution v ideo stills from m peg sequences [ J]. Proc Im age Processing, 1998, 14( 2): 465- 469.
[ 8] Sega ll C A, Katsagg elos A A, Rafae lM o lina, et al. Bayesian reso lution enhancem ent of compressed v ideo [ J] . IEEE Trans on Im age Process ing, 2004, 13( 7): 898- 911. [ 9] SegallC A, K atsagge los A A, Ra faelM olina, et al. Super - reso lution from compressed v ideo [ J]. Super-Reso lution Im ag ing, 2001, 19( 3): 211- 242

备注/Memo

备注/Memo:
基金项目: 江苏省自然科学基金资助项目( BK2004151)和安徽省高等学校自然科学基金资助项目( 2005K J075) .
作者简介: 周亮( 1981-) , 硕士研究生, 主要从事图像处理和多媒体通信的学习与研究. E-m ail: zk ing- zll@126. com
更新日期/Last Update: 2013-04-29