Non-linear Image Recovery trom a Single Frame Super Resolution Using Pearson Type VII Density

摘要

In this paper we formulated two versions of Pearson-MRF image priors, and conducted a comparative experimental study among these and state of the art methods of image prior from a single noisy version of low resolution image. We demonstrate that our proposed prior, univariate Pearson Type VII-MRF is likewise superior with Huber for all level of noise. The recovered image is always consistent although it has several local optima and we asses two different images.