Double adaptive filtering of Gaussian noise degraded images
SCIA'07 Proceedings of the 15th Scandinavian conference on Image analysis
Supervised restoration of degraded medical images using multiple-point geostatistics
Computer Methods and Programs in Biomedicine
| Hi-index | 0.00 |
The regularization of the least-squares criterion has been established as an effective approach of solving illposed image restoration problems. Unfortunately, a proper global regularization parameter is very difficult to be determined, and edges are usually smoothed by restoration process. In this paper, a new iterative regularization algorithm is presented. Before restoration, we divide the pixels of the blurred and noisy image into two types of regions: flat regions and edge regions (edges and the regions near edges). A non-local adaptive regularization function is used instead of a global regularization parameter, and a local regularization operator which is determined by the orientation of pixels is employed in edge regions. Experiments show that our algorithm is effective and the edge details are well preserved during the restoration process.