Multi-Grid Back-Projection Networks

TL;DR

The paper introduces Multi-Grid Back-Projection Networks, a multiscale convolutional architecture inspired by PDE solvers and iterative back-projection, for image and video super-resolution with efficient scaling and perceptual quality control.

Contribution

It proposes a novel multiscale network architecture that scales efficiently with resolution and introduces a noise-based strategy for controlling perceptual realism in super-resolution outputs.

Findings

Performs on par with state-of-the-art methods for exact high-resolution recovery.

Uses noise inputs to effectively control artificial detail generation.

Demonstrates improved perceptual quality through benchmark evaluations.

Abstract

Multi-Grid Back-Projection (MGBP) is a fully-convolutional network architecture that can learn to restore images and videos with upscaling artifacts. Using the same strategy of multi-grid partial differential equation (PDE) solvers this multiscale architecture scales computational complexity efficiently with increasing output resolutions. The basic processing block is inspired in the iterative back-projection (IBP) algorithm and constitutes a type of cross-scale residual block with feedback from low resolution references. The architecture performs in par with state-of-the-arts alternatives for regression targets that aim to recover an exact copy of a high resolution image or video from which only a downscale image is known. A perceptual quality target aims to create more realistic outputs by introducing artificial changes that can be different from a high resolution original content as long as they are consistent with the low resolution input. For this target we propose a strategy using noise inputs in different resolution scales to control the amount of artificial details generated in the output. The noise input controls the amount of innovation that the network uses to create artificial realistic details. The effectiveness of this strategy is shown in benchmarks and it is explained as a particular strategy to traverse the perception-distortion plane.