engleski

Non-linear image noise filtering algorithm based on SVD block processing

A non-linear algorithm for image noise filtering, based on singular value decomposition (SVD) is presented in this paper. Changes in singular values and singular vectors caused by additive white Gaussian noise are described. The non-linear filtering algorithm is based on eliminating that changes in singular values and singular vectors. Noise variance knowledge is not required. Value proportional to noise variance is estimated in the first phase of the algorithm, using least singular values. Once computed SVD of image blocks is then used for filtering operations. Processing of image in smaller blocks makes SVD-based procedure computationally feasible. Experimental results on various images have demonstrated the validity of the approach

image processing; singular value decomposition; image restoration

nije evidentirano