Discover layered structure in ultrasound images with a joint sparse representation model

Abstract

Deconvolution can improve the resolution of ultrasound system since the effect of point spread function can be removed. Traditional methods carry out the deconvolution line by line, so the neighborhood information are unused. A joint sparse model is proposed in this work to discover layered structure in ultrasound images. The model joints the axial deconvolution with a sparse lateral constraint. The model was tested on simulation and real data, and results support the enhanced performance in terms of resolution.