In this paper, a novel approach is proposed to remove the motion blur from a video, which is degraded and distorted by fast camera motion. Our approach is based on the image statistics rather than the traditional motion estimation. The image statistics has been successfully applied for blind motion deblurring for a single image by Fergus et al [3] and Levin [10]. Here a three-stage method is used to deal with the video. First, the “unblurred” frames in the video can be found based on the image statistics. Then the blur functions can be obtained by comparing the blurred frames with the unblurred ones. Finally a standard deconvolution algorithm is used to reconstruct the video. Our experiments show that our algorithms are efficient.
Three-Stage Motion Deblurring from a Video
Asian Conference on Computer Vision ; 2007 ; Tokyo, Japan November 18, 2007 - November 22, 2007
2007-01-01
10 pages
Article/Chapter (Book)
Electronic Resource
English
Motion Estimation , Motion Blur , Deconvolution Algorithm , Blur Kernel , Image Deconvolution Computer Science , Image Processing and Computer Vision , Computer Imaging, Vision, Pattern Recognition and Graphics , Pattern Recognition , Artificial Intelligence , Biometrics , Algorithm Analysis and Problem Complexity
Motion deblurring using hybrid imaging
IEEE | 2003
|Motion Deblurring Using Hybrid Imaging
British Library Conference Proceedings | 2003
|Motion Deblurring Using Non-stationary Image Modeling
British Library Online Contents | 2015
|NTRS | 1966
|Blind Motion Image Deblurring Based on Optimized Convolution Kernel
British Library Online Contents | 2014
|