Morphologic gain-controlled regularization for edge-preserving super-resolution image reconstruction

Total Variation or Bilateral Total variation-based regularization of ill-posed super-resolution (SR) problem is well established. However, the SR image reconstructed by this method produces ringing artifacts near strong edges. Second, the extension of SR Imaging to SR video always desire faster SR reconstruction process. We develop a gain-controlled-based locally adaptive regularization technique for SR reconstruction for faster convergence and more detail reconstruction while suppressing the ringing artifacts. We present an iterative process for the model and perform a series of numerical experiments to show evidence of the good performance of the numerical scheme and the proposed gain-controlled regularization.