Abstract

This study focuses on edge preservation and noise smoothing in the process of denoising. To achieve these two aims, the image has to be processed in such a way that the noise is removed to give people a pleasing vision without reducing the perceptibility of edges and details. In the proposed algorithm, edge orientations are taken into account by using directional templates during edge extraction. The results of convolution are adopted to control the coefficients of the corresponding denoising filter. These two steps play a leading role in guaranteeing the preservation of edges. Finally, flat regions distinguished from edges and details by local standard deviation are further operated on by incorporating the preliminary filtering result and mean filtering result to achieve better vision perception. Its great performance with lower complexity is validated by the experiential results, which provides an important opportunity for hardware implementation.

© 2020 Optical Society of America

A unified iterative denoising algorithm based on natural image statistical models: derivation and examples

Shan Tan and Licheng Jiao
Opt. Express 16(2) 975-992 (2008)

Wavelet-based noise-model driven denoising algorithm for differential phase contrast mammography

Carolina Arboleda, Zhentian Wang, and Marco Stampanoni
Opt. Express 21(9) 10572-10589 (2013)

Denoising algorithm of OCT images via sparse representation based on noise estimation and global dictionary

Xi Zhang, Zhongliang Li, Nan Nan, and Xiangzhao Wang
Opt. Express 30(4) 5788-5802 (2022)

Statistical model for OCT image denoising

Muxingzi Li, Ramzi Idoughi, Biswarup Choudhury, and Wolfgang Heidrich
Biomed. Opt. Express 8(9) 3903-3917 (2017)

BM3D adaptive TV filtering-based convolutional neural network for ESPI image denoising

Huamei Xin, Zengzhao Sun, Ying Xing, and Jingjing Wang
Appl. Opt. 60(35) 10920-10927 (2021)