Complementary Ensemble Empirical Mode Decomposition Based Microwave Induced Thermoacoustic Image Reconstruction Method

Abstract

In this paper, we adopt a signal processing method based on complementary ensemble empirical mode decomposition (CEEMD) and singular value decomposition (SVD) to reconstruct the thermoacoustic image. Thermoacoustic signals are easily interfered by factors such as temperature and mixed with incoherent noise during propagation, both CEEMD and SVD have a good effect in extracting the main components of the signal and removing noise. The main idea of this method is denoising artifacts by decomposing the ultrasound signal received by the sensor into a series of intrinsic mode functions (IMFs), choosing the effective IMFs based on SVD. We tested a single tumor in the homogeneous media by numerical simulation. The peak signal-to-noise ratios of the thermoacoustic images reconstructed by the proposed method, and the other three methods are compared. The results indicate that the method of CEEMD combined with SVD has better performance. Validation based on experimental data will be carried out in the follow-up work.