Blind Image Deconvolution Assisted by Robust Non-Gaussian Priors

摘要：

　　We propose a regularized blind image deconvolution method under Bayesian probabilistic framework.A prior distribution named Gaussian Scale Mixture Fields of Experts (GSM FOE) is adopted to regularize the original image and a sparse prior distribution is used to regularize the point spread function (PSF), the formulated problem is solved with an alternating minimization scheme.Experimental results on both synthetic and real world examples show that the proposed method is effective in suppressing negative artifacts which accompany blind image deconvolution and the restored images are of high quality.

关键词： blind image deconvolution GSM FoE Ip-norm regularization MAP estimation

作者: Shuyin Tao Wende Dong Zhenmin Tang Zhihai Xu

作者单位: Key Laboratory of Image and Video Understanding for Social Safety, Nanjing University of Science and The 28th Research Institute of China Electronic Technology Group Corporation, Nanjing, China State Key Laboratory of Modern Optical Instrumentation, Zhejiang University, Hangzhou, China

会议类型: 国际会议

会议名称: International Conference on Computational Science and Engineering Applications(CSEA2015)2015计算机科学与工程应用国际会议

会议地点: 三亚

会议语种:英文

页码: 84-91

在线出版日期: 2015-12-26（万方平台首次上网日期，不代表论文的发表时间）