Image Denoising based on Sparse Representation and Dual Dictionary
Aneesh G Nath1, Sreeram G2, Sharafudeen K3, Sreeraj M C4
1Aneesh G Nath, Department of CSE, TKM College of Engg., Kollam (Kerala), India.
2Sreeram G, Student, Department of CSE, TKM College of Engg., Kollam (Kerala), India.
3Sharafudeen K, Student, Department of CSE, TKM College of Engg.,Kollam, Kerala, India.
4Sreeraj M C Student, Department of CSE, TKM College of Engg., Kollam (Kerala), India.
Manuscript received on 15 April 2015 | Revised Manuscript received on 25 April 2015 | Manuscript Published on 30 April 2015 | PP: 249-253 | Volume-4 Issue-4, April 2015 | Retrieval Number: D3973044415/15©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Learning-based image denoising aims to reconstruct a denoised image from the prior model trained by a set of noised image patches. In this paper, we address the image denoising problem, where zero-mean white and homogeneous Gaussian additive noise is to be removed from a given image. Image denoising method via dual-dictionary learning and sparse representation consists of the main dictionary learning and the residual dictionary learning to recover denoised image. The approach taken is based on sparse representations over trained dictionaries. Using the K-SVD algorithm, we obtain a dictionary that describes the image content effectively. Using the corrupted or noised image primary main dictionary training is done. Since the K-SVD is limited in handling small image patches, we extend its deployment to arbitrary image sizes by defining a global image prior that forces sparsity over patches in every location in the image. We provide a residual dictionary learning phase which leads to a simple and effective denoising mechanism. This leads to a better denoising performance, and surpassing recently published leading alternative denoising methods. Extensive experimental results on test images validate that by employing the proposed two-layer progressive scheme, more image details can be recovered and much better results can be achieved in terms of both PSNR and visual perception.
Keywords: Sparse Representation, Dictionary Learning, Image Denoising, K-SVD, Residual Dictionary.
Scope of the Article: Deep Learning