Image Denoising Through Support Vector Regression

In this paper, an example-based image denoising algorithm is introduced. Image denoising is formulated as a regression problem, which is then solved using support vector regression (SVR). Using noisy images as training sets, SVR models are developed. The models can then be used to denoise different images corrupted by random noise at different levels. Initial experiments show that SVR can achieve a higher peak signal-to-noise ratio (PSNR) than the multiple wavelet domain Besov ball projection method on document images.

[1]  Edward R. Vrscay,et al.  Fractal image denoising , 2003, IEEE Trans. Image Process..

[2]  Jeffrey Ng,et al.  A steerable complex wavelet construction and its application to image denoising , 2005, IEEE Transactions on Image Processing.

[3]  Martin Vetterli,et al.  Spatially adaptive wavelet thresholding with context modeling for image denoising , 1998, Proceedings 1998 International Conference on Image Processing. ICIP98 (Cat. No.98CB36269).

[4]  Xue-Cheng Tai,et al.  Noise removal using smoothed normals and surface fitting , 2004, IEEE Transactions on Image Processing.

[5]  Richard G. Baraniuk,et al.  Improved wavelet denoising via empirical Wiener filtering , 1997, Optics & Photonics.

[6]  Martin J. Wainwright,et al.  Image denoising using scale mixtures of Gaussians in the wavelet domain , 2003, IEEE Trans. Image Process..

[7]  Narendra Ahuja,et al.  Image denoising using multiple compaction domains , 1998, Proceedings of the 1998 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP '98 (Cat. No.98CH36181).

[8]  Ruola Ning,et al.  Image denoising based on wavelets and multifractals for singularity detection , 2005, IEEE Transactions on Image Processing.

[9]  Richard G. Baraniuk,et al.  Multiple wavelet basis image denoising using Besov ball projections , 2004, IEEE Signal Processing Letters.

[10]  Vladimir N. Vapnik,et al.  The Nature of Statistical Learning Theory , 2000, Statistics for Engineering and Information Science.