Perception-Distortion Trade-off in the SR Space Spanned by Flow Models

TL;DR

This paper introduces an ensembling and fusion method for flow-based super-resolution models to improve the perception-distortion trade-off, reducing artifacts and enhancing fidelity while maintaining perceptual quality.

Contribution

It proposes a novel image fusion approach that leverages the diverse SR solutions from flow models to better balance perceptual quality and fidelity.

Findings

Improved perception-distortion trade-off over existing flow and adversarial models.

Effective reduction of visual artifacts in super-resolved images.

Enhanced quantitative and visual quality in experimental results.

Abstract

Flow-based generative super-resolution (SR) models learn to produce a diverse set of feasible SR solutions, called the SR space. Diversity of SR solutions increases with the temperature ($\tau$) of latent variables, which introduces random variations of texture among sample solutions, resulting in visual artifacts and low fidelity. In this paper, we present a simple but effective image ensembling/fusion approach to obtain a single SR image eliminating random artifacts and improving fidelity without significantly compromising perceptual quality. We achieve this by benefiting from a diverse set of feasible photo-realistic solutions in the SR space spanned by flow models. We propose different image ensembling and fusion strategies which offer multiple paths to move sample solutions in the SR space to more desired destinations in the perception-distortion plane in a controllable manner depending on the fidelity vs. perceptual quality requirements of the task at hand. Experimental results demonstrate that our image ensembling/fusion strategy achieves more promising perception-distortion trade-off compared to sample SR images produced by flow models and adversarially trained models in terms of both quantitative metrics and visual quality.