Image Denoising using a Combination of Spatial Domain Filters and Convolutional Neural Networks
Rajeshkannan Regunathan1, Punith N S2, Ashraf Ali K3, Gautham S4
1Rajeshkannan Regunathan,School of Computer Science and Engineering, VIT University, Vellore, (Tamil Nadu), India.
2Punith NS, School of Computer Science and Engineering, VIT University, Vellore, (Tamil Nadu), India.
3Ashraf Ali K,School of Computer Science and Engineering, VIT University, Vellore, (Tamil Nadu), India.
4Gautham S,School of Computer Science and Engineering, VIT University, Vellore, (Tamil Nadu), India.
Manuscript received on 23 March 2019 | Revised Manuscript received on 30 March 2019 | Manuscript published on 30 March 2019 | PP: 1836-1841 | Volume-7 Issue-6, March 2019 | Retrieval Number: F2810037619/19©BEIESP
Open Access | Ethics and 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: Image denoising is one of the most pressing challenges in Image Processing. Spatial Domain Filters are long existing method for denoising that are simple and effective against different types of noises, but they do not take into account the recurring patterns. Furthermore, due to ever increasing demand for image denoising on various applications in technology, the computational and memory intensiveness along with their performance on various types of noises becomes extremely important. Recently Convolutional Neural Networks have turned out to be state-of-the-art methods for denoising. We put forward a system which integrates a deep CNN preceded by Block Matching along with traditional spatial domain methods for image denoising for both random structures of an image as well as recurring patterns. This system is evaluated over a large data set of grey-scale images and various noises and has state-of-theart results.
Keywords: Image denoising, Spatial domain filters, Convolutional neural networks, Block matching
Scope of the Article: Image Security