Perception-Distortion Trade-off in the SR Space Spanned by Flow Models
This work addresses the problem of balancing visual quality and accuracy in image super-resolution for applications like photography or medical imaging, but it is incremental as it builds on existing flow-based models.
The paper tackles the trade-off between perceptual quality and distortion in super-resolution (SR) by proposing an image ensembling/fusion approach to reduce artifacts and improve fidelity in flow-based SR models, achieving a more promising perception-distortion trade-off compared to existing methods.
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 ($τ$) 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.