3D MR Images Denoising using Adaptive Block wise Approached Non-Local Means (ABNLM) Filter for Spatially Varying Noise Levels
Poornaiah Billa1, Anandbabu Gopatoti2
1Poornaiah Billa, Department of Electronics and Instrumentation Engineering, Lakireddy Bali Reddy College of Engineering, Mylavaram ( Andhra Pradesh), India.
2Anandbabu Gopatoti, Department of Electronics and Communication Engineering, MVR College of Engineering & Technology, Paritala (Andhra Pradesh), India.
Manuscript received on 05 February 2019 | Revised Manuscript received on 13 February 2019 | Manuscript published on 28 February 2019 | PP: 112-118 | Volume-8 Issue-4, February 2019 | Retrieval Number: D2658028419/19©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: The uniform noise distribution over the image is assumed in most of the filtering techniques. The resulting filtering technique becomes problematic when noise not uniformly distributed. Magnetic Resonance images with spatially varying noise levels were produced by Sensitivity-encoded, intensity inhomogeneity and surface coil based acquisition techniques. To adapt these spatial variations in noise levels, we propose a new Adaptive Blockwise approached NL-Means Filter where denoising capability of filter is adjusted based on the local image noise level. Image Noise levels are spontaneously acquired from the MR images using a proposed new adaptive technique. To reduce the computational burden of NLM Filter, an Adaptive Blockwise Non-Local Means Filter is proposed to speed up the denoising process. With adaptive soft wavelet coefficient mixing, a multiresolution framework is adapted to ABNLM filter for denoising of 3-Dimensional MR images. The proposed Multiresolution filter adapts the filtering parameters automatically based on image space-frequency resolution. The outcome of the stated multiresolution Adaptive Blockwise Non-Local Means Filter shows better performance in considering the non uniform noise when compared to Rician NL-means filters where the noise parameters has to be specified initially.
Keyword: Non-Local Mean Filter, Blockwise Approach, Magnetic Resonance (MR) Image, Wavelet Transform and Denoising.
Scope of the Article: 3D Printing