Image interpolation using a Bayesian approach in wavelet domain

The goal of medical image resolution enhancement is to reconstruct a higher-resolution image from its lower-resolution counterpart. This paper proposes a Bayesian approach in the wavelet domain by exploiting a Bayesian inference framework to mathematically formulate the image interpolation problem. Furthermore, the proposed approach jointly estimates both the unknown wavelet coefficients of the high-resolution image and the unknown parameters of the statistical model for wavelet coefficients. Experiments are conducted to demonstrate the superior performance of the proposed approach.