Variational image deringing using varying regularization parameter
I. T. Sitdikov, Andrey Serdjevich Krylov · Pattern Recognition and Image Analysis · 2015
In this paper we propose a new variational image deringing method. This method is based on total variation minimization with an adaptively varying regularization parameter. The proposed approach improves visual quality of resulting images by preserving more structural information comparing to existing methods. Attendant parallel algorithms have been developed and implemented on GPU.