Adaptive space-variant single image deblurring method based on saliency map

被引：0
|
作者：
Shi Y. [1 ,2 ]
Yan J. [1 ,2 ]
Huang Z. [1 ,2 ]
Hua X. [1 ,2 ]
机构：
[1] School of Electrical and Information Engineering, Wuhan Institute of Technology, Wuhan
[2] Recognition, Wuhan Institute of Technology, Wuhan
[3] Laboratory of Hubei Province Video Image and HD Projection Engineering Technology, Research Center, Wuhan Institute of Technology, Wuhan
来源：
Huazhong Keji Daxue Xuebao (Ziran Kexue Ban)/Journal of Huazhong University of Science and Technology (Natural Science Edition) | 2021年 / 49卷 / 09期
关键词：
Adaptive deblurring; Image dividing; Quadtree decomposition; Saliency map; Weighted window function;
D O I：
10.13245/j.hust.210906
中图分类号：
学科分类号：
摘要：
The irrationality of image dividing will lead to the problems of the inaccurate kernel estimation and excessive computation. In order to solve this problem, an efficient method for adaptively recovering a single spatially variant blurred image was proposed. Based on the blur kernels' similarity to ensure that the obtained patches have different kernels, the quadtree decomposition was introduced to adaptively partition the image. Then a saliency map was added to recover the latent image of each patch, which is capable extracting the essential parts that human eyes are interested to estimate kernel more accurately and preserve more image details. Finally, a weighted window function was combined to joint adjacent patches with different sizes to get the deblurring result. The experimental results on both space-invariant and space-variant images show that our proposed method can achieve the comparable or better restoration compared with the existing classic methods. © 2021, Editorial Board of Journal of Huazhong University of Science and Technology. All right reserved.
引用
收藏
页码：30 / 35
页数：5
相关论文
共 18 条
  • [1] ZHU X, MILANFAR P., Removing atmospheric turbulence via space-invariant deconvolution, IEEE Transaction on Pattern Analysis and Machine Intelligence, 35, 1, pp. 157-170, (2013)
  • [2] XIE Y, ZHANG W, TAO D, Et al., Removing turbulence effect via hybrid total variation and deformation-guided kernel regression, IEEE Transactions on Image Processing, 25, 10, pp. 4943-4958, (2016)
  • [3] 42, 3, pp. 28-32, (2014)
  • [4] 3, pp. 50-53
  • [5] XU L, ZHENG S, JIA J., Unnatural L<sub>0</sub> sparse representation for natural image deblurring, Proc of Conference on Computer Vision and Pattern Recognition (CVPR), pp. 1107-1114, (2013)
  • [6] HU Z, XU L., Joint depth estimation and camera shake removal from single blurry image, Proc of Conference on Computer Vision and Pattern Recognition (CVPR), pp. 2893-2900, (2014)
  • [7] COSTELLO T P, MIKHAEL W B., Efficient restoration of space-variant blurs from physical optics by sectioning with modified wiener filtering, Digital Signal Processing, 13, 1, pp. 1-22, (2003)
  • [8] LEVIN A., Blind motion deblurring using image statistics, Advances in Neural Information Processing Systems, 2, pp. 1738-1742, (2006)
  • [9] JOSHI N, KANG S B, ZITNICK C L, Et al., Image deblurring using inertial measurement sensors, ACM Transactions on Graphics, 29, 4, pp. 1-30, (2010)
  • [10] WHYTE O, SIVIC J, ZISSERMAN A, Et al., Non-uniform deblurring for shaken images, Proc of Conference on Computer Vision and Pattern Recognition (CVPR), pp. 491-498, (2012)
12