Single Image Super-Resolution via L 0 Image Smoothing

Abstract EN

We propose a single image super-resolution method based on a L 0 smoothing approach.

We consider a low-resolution image as two parts: oneis the smooth image generated by the L 0 smoothing method and the other isthe error image between the low-resolution image and the smoothing image.

Weget an intermediate high-resolution image via a classical interpolation and thengenerate a high-resolution smoothing image with sharp edges by the L 0 smoothingmethod.

For the error image, a learning-based super-resolution approach, keepingimage details well, is employed to obtain a high-resolution error image.

Theresulting high-resolution image is the sum of the high-resolution smoothing imageand the high-resolution error image.

Experimental results show the effectivenessof the proposed method.