Training a Task-Specific Image Reconstruction Loss

TL;DR

This paper introduces a set of application-specific loss functions trained with discriminators to improve image restoration tasks, outperforming traditional perceptual losses like VGG in super resolution, denoising, and JPEG artifact removal.

Contribution

It proposes a novel, application-specific loss training method using discriminators, eliminating the need for large pre-trained networks like VGG.

Findings

Outperforms state-of-the-art loss functions in super resolution, denoising, and JPEG artifact removal.

Requires only a single natural image and its distortions for training.

Specialized loss functions effectively identify distortions rather than predict perceived quality.

Abstract

The choice of a loss function is an important factor when training neural networks for image restoration problems, such as single image super resolution. The loss function should encourage natural and perceptually pleasing results. A popular choice for a loss is a pre-trained network, such as VGG, which is used as a feature extractor for computing the difference between restored and reference images. However, such an approach has multiple drawbacks: it is computationally expensive, requires regularization and hyper-parameter tuning, and involves a large network trained on an unrelated task. Furthermore, it has been observed that there is no single loss function that works best across all applications and across different datasets. In this work, we instead propose to train a set of loss functions that are application specific in nature. Our loss function comprises a series of discriminators that are trained to detect and penalize the presence of application-specific artifacts. We show that a single natural image and corresponding distortions are sufficient to train our feature extractor that outperforms state-of-the-art loss functions in applications like single image super resolution, denoising, and JPEG artifact removal. Finally, we conclude that an effective loss function does not have to be a good predictor of perceived image quality, but instead needs to be specialized in identifying the distortions for a given restoration method.