ASDN: A Deep Convolutional Network for Arbitrary Scale Image Super-Resolution

TL;DR

This paper introduces ASDN, a deep convolutional network utilizing a Laplacian pyramid approach to efficiently perform super-resolution at arbitrary scales, outperforming existing methods in PSNR and computational cost.

Contribution

The paper proposes a novel Laplacian pyramid-based deep network for arbitrary scale super-resolution, reducing computational costs and improving performance over fixed-scale methods.

Findings

ASDN outperforms fixed-scale methods like SRCNN, VDSR, DRRN by about 1 dB in PSNR.

ASDN exceeds Meta-SR on many scales in arbitrary scale super-resolution.

The recursive approach significantly reduces computational costs for larger scales.

Abstract

Deep convolutional neural networks have significantly improved the peak signal-to-noise ratio of SuperResolution (SR). However, image viewer applications commonly allow users to zoom the images to arbitrary magnification scales, thus far imposing a large number of required training scales at a tremendous computational cost. To obtain a more computationally efficient model for arbitrary scale SR, this paper employs a Laplacian pyramid method to reconstruct any-scale high-resolution (HR) images using the high-frequency image details in a Laplacian Frequency Representation. For SR of small-scales (between 1 and 2), images are constructed by interpolation from a sparse set of precalculated Laplacian pyramid levels. SR of larger scales is computed by recursion from small scales, which significantly reduces the computational cost. For a full comparison, fixed- and any-scale experiments are conducted using various benchmarks. At fixed scales, ASDN outperforms predefined upsampling methods (e.g., SRCNN, VDSR, DRRN) by about 1 dB in PSNR. At any-scale, ASDN generally exceeds Meta-SR on many scales.