Ultrasound image deconvolution based on dynamic SNR estimation

Compared with other medical imaging modes, ultrasound imaging suffers from its poor resolution caused by the signal convolution effect. In order to solve the problem, a deconvolution processing should be performed. Wiener filtering is a conventional and effective deconvolution algorithm, and the Signal-to-Noise Ratio (SNR) parameter plays an important effect on the deconvolution result. In ultrasound imaging, the image SNR varies with the depth, so a segmented Wiener deconvolution method is adopted in this paper. Firstly, the ultrasound image is segmented to several zones along the depth direction. Then the local SNR of each segment image is estimated. Finally, the deconvolution is performed within the segment image based on the estimated local SNR. The simulation experiment results indicate that ultrasound imaging resolution is improved more effectively than the traditional deconvolution with a fixed SNR.