In this paper, we present an improved algorithm for super-resolution image restoration which used L-Curve regularization to Maximum A-Posteriori method. Super-Resolution as the second-generation problem of image restoration is also an ill-posed question because of an insufficient number of LR images and ill-conditioned blur operators. Procedures adopted to stabilize the inversion of ill-posed problem are called regularization, so the selection of regularization parameter is very important to the effect of image reconstruction. Hence, we present L-Curve regularization method which can estimate the regularization parameter exactly and not by trial-and-error, then used the regularization parameter to Maximum A-Posteriori method which is one of the most common methods of super-resolution image restoration. Experiments demonstrate that this method can effectively reduce and remove noise in the restoration images, get good quality restoration images and has high super resolution performance.
Super-Resolution Image Restoration with L-Curve
2008 Congress on Image and Signal Processing ; 3 ; 597-601
2008-05-01
316131 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Super-Resolution Image Restoration from Blurred Low-Resolution Images
British Library Online Contents | 2005
|An empirical analysis of super resolution techniques for image restoration
DSpace@MIT | 2000
|MPMAP super-resolution image restoration algorithm based on multiframes or multisensors [4925-02]
British Library Conference Proceedings | 2002
|British Library Online Contents | 2006
|