A Denoising Module Based On the Wavelet and Empirical Mode Decomposition for Photoacoustic Microscopy

In this work, we set up a denoising module to improve the imaging result for the photoacoustic microscopy (PAM) by improving the signal noise ratio. This module contains a series of data processing methods to reduce the noise from the tissues and the system. This module is adaptive to different imaging systems because of these methods' intrinsic characteristics. Meanwhile, the parameters are decided based on the property of data being denoised. In this module, firstly data length is limited because each system has its own imaging depth capacity and data outside is mostly noise. Data is filtered in frequency domain in accordance with bandwidth of the imaging system and the image is filtered with the Wiener adaptive filter. Secondly data is presented in time-frequency domain with different time-frequency analysis methods. With the aid of this presentation in time-frequency domain, we can decide denoising parameters based on the characteristics of denoised data. Thirdly data is denoised using wavelet and empirical mode decomposition (EMD) methods. These methods demonstrate strong denoising capacity in the data processing field and are very suitable for processing data from biological tissues. With decided parameters, wavelet and EMD methods are set down and data is denoised automatically to get the best imaging effect instead of processing each data manually with different methods. This denoising module improves the imaging quality and has adaptive ability to reach promising result for different PAM imaging systems.