Abstract | ||
---|---|---|
Image denoising is a classical linear inverse problem with applications in remote sensing, medical imaging, astronomy and surveillance. This article addresses the image denoising problem using a non-local noise estimation based on the spatial redundancy offered by natural images. A low dimensional signal subspace is estimated using the statistical strength of singular value decomposition (SVD), which reduces the computational burden and enhances the local basis screening. A multiple regression based approach is then applied on the estimated basis to calculate the observation noise and the whole image is restored by patch based processing. The proposed method is adaptive in the sense that all the algorithm parameters are learned from the observed noisy data. The simulated comparisons shows comparatively high performance of the proposed algorithm comparing to the other image denoising techniques. |
Year | DOI | Venue |
---|---|---|
2015 | 10.1109/DICTA.2015.7371290 | 2015 International Conference on Digital Image Computing: Techniques and Applications (DICTA) |
Keywords | Field | DocType |
nonlocal noise estimation,adaptive image denoising,classical linear inverse problem,remote sensing,medical imaging,astronomy,surveillance,image denoising problem,spatial redundancy,natural image,signal subspace,statistical strength,singular value decomposition,SVD,local basis screening,multiple regression based approach,observation noise,patch based processing,algorithm parameter,image denoising technique | Noise reduction,Computer vision,Singular value decomposition,Pattern recognition,Basis pursuit denoising,Noise measurement,Non-local means,Computer science,Artificial intelligence,Image restoration,Signal subspace,Video denoising | Conference |
Citations | PageRank | References |
0 | 0.34 | 22 |
Authors | ||
2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Muhammad Hanif | 1 | 207 | 25.54 |
Abd-Krim Seghouane | 2 | 193 | 24.99 |