基于广义规范化稀疏模型的图像盲去模糊算法
针对基于规范化稀疏先验的图像盲去模糊方法估计精度低、计算速度慢、参数选择敏感等问题,提出一种Tikhonov正则增强的广义规范化稀疏模型,且将其作为中间清晰图像和运动模糊核的共同先验约束。随后,利用算子分裂、交替方向乘子法以及快速傅里叶变换,最小化关于中间清晰图像与运动模糊核的目标函数,导出一种快速图像盲去模糊算法。在标准测试集以及实际彩色模糊图像上的实验结果验证了提出方法的有效性和鲁棒性;此外,在同等条件下与近期文献中的盲去模糊方法进行比较,显示了该方法在估计精度和估计效率上的双重优势。...
Saved in:
| Published in | 计算机应用研究 Vol. 34; no. 8; pp. 2533 - 2537 |
|---|---|
| Main Author | |
| Format | Journal Article |
| Language | Chinese |
| Published |
南京邮电大学 通信与信息工程学院,南京,210003
2017
|
| Subjects | |
| Online Access | Get full text |
| ISSN | 1001-3695 |
| DOI | 10.3969/j.issn.1001-3695.2017.08.066 |
Cover
| Summary: | 针对基于规范化稀疏先验的图像盲去模糊方法估计精度低、计算速度慢、参数选择敏感等问题,提出一种Tikhonov正则增强的广义规范化稀疏模型,且将其作为中间清晰图像和运动模糊核的共同先验约束。随后,利用算子分裂、交替方向乘子法以及快速傅里叶变换,最小化关于中间清晰图像与运动模糊核的目标函数,导出一种快速图像盲去模糊算法。在标准测试集以及实际彩色模糊图像上的实验结果验证了提出方法的有效性和鲁棒性;此外,在同等条件下与近期文献中的盲去模糊方法进行比较,显示了该方法在估计精度和估计效率上的双重优势。 |
|---|---|
| Bibliography: | 51-1196/TP blind image deblurring; normalized sparsity prior; ringing artifacts; ADMM Yang Changxing, Shao Wenze, Ge Qi, Li Haibo （ College of Telecom munications ＆ Information Engineering, Nanjing University of Posts ＆ Telecommunications, Nanjing 210003, China） This paper proposed a new blind deblurring method utilizing the generalized normalized sparsity （NS） prior as a common constraint on both the intermediate shaW image and motion blur kernel, which was boosted by the well-known Tik- honov regularization. The motivation of the new approach was that, despite NS empirically showed to prefer shaW images rather than blurred ones, the NS-based blind deblurring method was actually of lower accuracy, higher computational cost and greater sensitivity to choice of parameters. Then, it combined the operator splitting, alternating direction method of multipliers （ADMM） , and fast Fourier transform to minimize the objective function with respect to the intermediate sharp image and mo- tion blur kernel, led to a fast bl |
| ISSN: | 1001-3695 |
| DOI: | 10.3969/j.issn.1001-3695.2017.08.066 |