Quality Improvement of Thermoacoustic Imaging Based on Compressive Sensing

Abstract

Thermoacoustic imaging (TAI) is a non-ionizing, high-contrast and high-resolution modality for biomedical applications. The resolution of the reconstructed image by traditional imaging algorithm, such as back-projection (BP), is dependent on the input microwave pulse width. In this work, it is experimentally demonstrated that the quality of images reconstructed based on compressive sensing (CS), a novel imaging algorithm that can reconstruct images using fewer measurements, is largely independent on the microwave pulse width which is taken into account when building the dictionary. Quantitative analysis reveals that the peak signal to noise ratio of the reconstructed images based on CS is at least 6 dB higher than those via BP, and the correlation coefficients of the images based on CS show higher correlation than those via BP.