[1]何富运,张志胜.基于马尔科夫随机场学习模型的图像模糊核估计[J].东南大学学报(自然科学版),2016,46(6):1143-1148.[doi:10.3969/j.issn.1001-0505.2016.06.006]
 He Fuyun,Zhang Zhisheng.Image blur kernel estimation based on Markov random field learning model[J].Journal of Southeast University (Natural Science Edition),2016,46(6):1143-1148.[doi:10.3969/j.issn.1001-0505.2016.06.006]
点击复制

基于马尔科夫随机场学习模型的图像模糊核估计()
分享到:

《东南大学学报(自然科学版)》[ISSN:1001-0505/CN:32-1178/N]

卷:
46
期数:
2016年第6期
页码:
1143-1148
栏目:
计算机科学与工程
出版日期:
2016-11-20

文章信息/Info

Title:
Image blur kernel estimation based on Markov random field learning model
作者:
何富运张志胜
东南大学机械工程学院, 南京 211189
Author(s):
He Fuyun Zhang Zhisheng
School of Mechanical Engineering, Southeast University, Nanjing 211189, China
关键词:
图像恢复 模糊核 马尔科夫随机场 核相似性
Keywords:
image restoration blur kernel Markov random field kernel similarity
分类号:
TP391
DOI:
10.3969/j.issn.1001-0505.2016.06.006
摘要:
为在图像模糊核估计中充分利用图像的区域特征和结构信息作为先验知识,提出一种基于马尔科夫随机场学习模型的模糊核估计方法.首先,由滑动的子窗口构成马尔科夫随机场的节点集,以每个子窗口的曲率方向能量滤波器的响应和边缘分布组成的特征向量作为模型的输入;然后,利用对数伪似然优化算法估计模型参数,在模型训练阶段,采用交叉熵相似性度量模糊核的相似性以标记训练样本;最后,利用置信度传播算法推测最优图像子块.运用所提方法对仿真和实际模糊图像进行实验, 结果表明,该学习模型可以精确地估计模糊核,在主观视觉对比和客观评价方面均具有较好的效果,同时也具有较好的自适应性.与其他3种方法相比,模糊核相似度分别提高了1.55%,5.64%和7.02%.
Abstract:
To make the most of image’s regional feature and structural information as the prior knowledge in estimating blur kernel, an estimation method for blur kernel based on the Markov random field learning model is proposed. First, a node set in the Markov random field is constituted by sliding sub-window, and the image characteristics of each sub-window, such as the response of multi-curvature orientation energy filter and edge distribution, are extracted as the input vector. Then, model parameters are estimated by the logarithmic pseudo-likelihood optimization algorithm, and the training samples are labeled by adopting the cross entropy similarity to measure blur kernel’s similarity. Finally, the optimal image sub-window is inferred based on the loopy belief propagation algorithm. Both synthetic and real blurred images are tested by the proposed method. The experimental results show that the method can accurately estimate blur kernel, and achieves favorable effects both in subjective visual contrast and objective evaluation. Meanwhile, the method also has a strong self-adaptability. Compared with the other three methods, the blur kernel similarity is improved by 1.55%, 5.64% and 7.02%, respectively.

参考文献/References:

[1] Babacan S D, Molina R, Do M N, et al. Bayesian blind deconvolution with general sparse image priors[C]//European Conference on Computer Vision. Florence, Italy,2012:341-355. DOI: 10.1007/978-3-642-33783-3-25.
[2] Fortunato H E, Oliveira M. Fast high-quality non-blind deconvolution using sparse adaptive priors [J]. Visual Computer, 2014, 30(6):661-671. DOI: 10.1007/s00371-014-0966-x.
[3] Ma J, Zhang Z B, Yang H, et al. A study on defocused iris image restoration [J]. Acta Automatica Sinica, 2012, 38(6):959-968. DOI:10.3724/sp.j.1004.2012.00959.
[4] Tikhonov A N, Arsenic V Y. Solution of ill-posed problem [M]. New York: Wiley, 1977:11-17.
[5] Kong Y, Wang X. An algorithm of blind restoration for defocus blurred image [J]. Application of Electronic Technique, 2013, 39(1):125-131. DOI:10.16157/j.issn.0258-7998.2013.01.048.
[6] Freeman W T, Fergus R D W, Singh B, et al. Removing camera shake from a single photograph: US Patent, US20080025627 [P]. 2008.
[7] Levin A, Weiss Y, Durand F, et al. Understanding and evaluating blind deconvolution algorithms[C]//2009 IEEE Conference on Computer Vision and Pattern Recognition. Miami, FL, USA, 2009:1964-1971. DOI:10.1109/cvprw.2009.5206815.
[8] Bae H, Fowlkes C, Chou P H. Patch mosaic for fast motion deblurring[C]// Asian Conference on Computer Vision. Daejeon, Korea, 2012:322-335. DOI: 10.1007/978-3-642-37431-9-25.
[9] Xu L, Jia J. Two-phase kernel estimation for robust motion deblurring[C]//European Conference on Computer Vision. Heraklion, Crete, Greece, 2010:81-84.
[10] Yedidia J S, Freeman W T, Weiss Y. Understanding belief propagation and its generalizations[J]. Exploring Artificial Intelligence in the New Millenium, 2003, 54(1):276-286.
[11] Bobadilla J, Ortega F, Hernando A. A collaborative filtering similarity measure based on singularities [J]. Information Processing & Management, 2012, 48(2):204-217. DOI:10.1016/j.ipm.2011.03.007.
[12] 张德平,查日军.基于Markov链使用模型的加速统计测试方法[J].东南大学学报(自然科学版),2011, 41(5):949-955.
  Zhang Deping, Zha Rijun. Acceleration statistical testing method based on Markov chain usage model [J]. Journal of Southeast University(Natural Science Edition), 2011, 41(5): 949-955.(in Chinese)
[13] Bong D B L, Khoo B E. Objective blur assessment based on contraction errors of local contrast maps [J]. Multimedia Tools & Applications, 2015, 74(17):7355-7378. DOI: 10.1007/s11042-014-1983-5.
[14] Bengio Y, Courville A, Vincent P. Representation learning: A review and new perspectives [J]. IEEE Transactions on Software Engineering, 2013, 35(8):1798-1828. DOI:10.1109/TPAMI.2013.50.
[15] Abbeel P, Koller D, Ng A Y. Learning factor graphs in polynomial time & sample complexity [J]. Journal of Machine Learning Research, 2012, 7(4):1743-1788.
[16] Hu Z, Yang M H. Learning good regions to deblur images [J]. International Journal of Computer Vision, 2015, 115(3):1-18. DOI: 10.1007/s11263-015-0821-1.
[17] Levin A, Weiss Y, Durand F, et al. Efficient marginal likelihood optimization in blind deconvolution[C]//IEEE Conference on Computer Vision & Pattern Recognition. Washington, DC, USA, 2011:2657-2664.
[18] Fergus R, Singh B, Hertzmann A, et al. Removing camera shake from a single photograph [J]. ACM Transactions on Graphics, 2006, 25(3):787-794. DOI:10.1145/1141911.1141956.

相似文献/References:

[1]栾中,尚媛园,舒华忠,等.基于SLIC和SVR的单幅图像去雾算法[J].东南大学学报(自然科学版),2018,48(1):25.[doi:10.3969/j.issn.1001-0505.2018.01.005]
 Luan Zhong,Shang Yuanyuan,Shu Huazhong,et al.Single image dehazing algorithm based on SLIC and SVR[J].Journal of Southeast University (Natural Science Edition),2018,48(6):25.[doi:10.3969/j.issn.1001-0505.2018.01.005]

备注/Memo

备注/Memo:
收稿日期: 2016-02-20.
作者简介: 何富运(1982—),男,博士生;张志胜(联系人),男,博士,教授,博士生导师,oldbc@seu.edu.cn.
基金项目: 国家自然科学基金资助项目(51275090)、国家自然科学基金科学仪器基础研究专款资助项目(21327007)、中央高校基本科研业务费专项资金资助项目、江苏省普通高校研究生科研创新计划资助项目(KYLX15_0208).
引用本文: 何富运,张志胜.基于马尔科夫随机场学习模型的图像模糊核估计[J].东南大学学报(自然科学版),2016,46(6):1143-1148. DOI:10.3969/j.issn.1001-0505.2016.06.006.
更新日期/Last Update: 2016-11-20