Download presentation
Presentation is loading. Please wait.
Published byDiego Goodier Modified over 10 years ago
1
Blind motion deblurring from a single image using sparse approximation
Jian-Feng Caiy, Hui Jiz, Chaoqiang Liuy and Zuowei Shenz National University of Singapore, Singapore Center for Wavelets, Approx. and Info. Proc.y and Department of Mathematicsz CVPR 2009 報告者:黃智勇 2017/4/6
2
Outline Introduction Tight framelet system and curvelet system
Sparse representation under framelet and curvelet system Formulation of our minimization Numerical algorithm and analysis Experiments 2017/4/6
3
Introduction We propose to use framelet system (Ron and Shen et al. [24]) to find the sparse approximation to the image under framelet domain. We use the curvelet system (Candes and Donoho [8]) to find the sparse approximation to the blur kernel under curvelet domain. 2017/4/6
4
Tight famelet system 2017/4/6
5
Sparse representation under framelet and curvelet system
2017/4/6
6
Formulation of our minimization
We denote the image g (or the kernel p) as a vector g (or p). Let “。” denote the usual 2D convolution after column concatenation, then we have Let u = Ag denote the framelet coefficients of the clear image g, and let v = Cp denote the curvelet coefficients of the blur kernel p. 2017/4/6
7
Numerical algorithm and analysis
there exist only two difficult problems (14) and (15) of the same type. For such a large-scale minimization problem with up to millions of variables, there exists a very efficient algorithm based on so-called linearized Bregman iteration technique. 2017/4/6
8
Numerical algorithm and analysis
2017/4/6
9
Experiments 2017/4/6
10
Experiments 2017/4/6
11
Experiments 2017/4/6
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.