Single image super-resolution method based on bilinear interpolation and U-Net combination

Abstract

The single image super-resolution issue is studied in this paper. We propose a new method based on a combination of bilinear interpolation and the U-Net neural network to solve this problem. The image is enlarged by bilinear interpolation, then its quality is improved by the neural network. This approach allows improving reconstruction quality by 1.00-6.56% compared with known methods