Image denoising using 2-D wavelet algorithm for Gaussian-corrupted confocal microscopy images

Confocal laser scanning microscopy (CLSM) imaging is a non-invasive optical imaging technique for the examination of the living tissues. CLSM inherently enables in-depth sectioning (z-slices) of the focused specimen. Z-slices of the targeted tissue are gathered by adjusting the focal point on the z-axis into the tissue. Unfortunately, these images can get corrupted with noise of different levels caused by out-of focus light originating from above and below the focal plane. This study proposes a reliable method to indicate and eliminate the additive white Gaussian noise (AWGN) present in real CLSM images of skin tissue. In this work, a denoising algorithm using discrete wavelet transform (DWT) is developed in order to remove the noise by preserving the energy conservation. The effect and performance of different wavelet thresholding algorithms are compared and studied along with different tuning parameters. The selection of components employed in the algorithm affects the noise reduction performance therefore, a systematic approach is presented to obtain and utilize the best combination of these parameter values. Analysis of variance (ANOVA) is exploited to inspect the main and the interaction effects of treated parameters. Computational results show the effectiveness of the methodical tuning approach to CLSM image denoising. (C) 2019 Published by Elsevier Ltd.