A two-stage synthetic non-local mean (NLM) filtering method for efficient denoising in quantitative phase imaging (QPI)

Abstract

AbstractTo improve the accuracy of phase recovery in quantitative phase imaging, a novel two-stage synthetic denoising method is proposed in this paper. In this method, to avoid the loss of effective detail information, the mode of fixed attenuation parameter setting is changed into two stages of self-adaptation setting: the first attenuation parameter is defined as the convolution of filtered pixels in the window of interest, the second attenuation parameter is defined as the noise standard deviation of each pixel after filtering in the first stage. Moreover, the Pearson coefficient expression is redefined to get a more accurate similarity evaluation, and the algorithm's operational efficiency has been improved using cross-correlation functions and fast Fourier transform. The comparative results of simulations and experiments show that this method is more effective in reducing speckle noise in quantitative phase imaging and enhancing phase recovery accuracy, providing a new digital image denoizing method for efficient and accurate imaging.KEYWORDS: Quantitative phase imagingefficient denoisingadaptivetwo stages process Disclosure statementNo potential conflict of interest was reported by the author(s).Additional informationFundingThis work was supported by the National Natural Science Foundation of China: [grant number No. 11874184].