Restoration of medical ultrasound images via modified homomorphic deconvolution

Compared with other medical imaging devices, ultrasound imaging suffers from its poor resolution. Therefore, in order to obtain the important diagnostic image detail, restoration procedure need to be performed. The paper describes a modified homomorphic deconvolution which is used to improve the quality of medical ultrasound image. The proposed deconvolution performs the homomorphic filtering based on the estimation of the point-spread function (PSF). Especially, the application of a non-local means (NL-means) algorithm makes PSF estimation more precise for rejecting the White-Gaussian noise (WGN) effectively. We validate our method for different radio-frequency (RF) images with resolution improvement.