Regularized Processing of Signal Deconvolution in Photo-Acoustic Signal Recovery

Zhiyuan Song,Yixiong Su,Tao Lu,Fan Zhang,Ruikang K. Wang

DOI: https://doi.org/10.1117/12.705739

2007-01-01

Abstract:In photoacoustic (PA) tomography, a piezoelectrical signal of inner characteristic of interesting object is mainly acquired by a hydrophone. Every piezoelectrical signal as output signal is the convolution of the original input signal that denotes the ultrasonic signal emitting from the substance and the system transfer function. The undistorted input signal is the very physical quantity that we want actually. Therefore an original input signal is computed with the deconvolution of the system transfer function and the output signal. While most practical deconvolution problems are called as blind deconvolution because the system transfer function and the input signal are both unknown and estimated from the output signal in the same time. In common, the deconvolution problem has an important property that it is called ill-condition, which is a special and intractable difficulty that both the theoretic analysis and the numerical computation would meet. For the sake of getting the solution of the deconvolution problem reasonable in physics and responsible for the gained data continually, a package of theory method called regularization to cure the ill-conditioned problems is applied in the PA signal processing.