Stop Throwing Away Discriminators! Re-using Adversaries for Test-Time Training

TL;DR

This paper proposes re-using trained GAN discriminators during test time to improve segmentation accuracy by fine-tuning the segmentor with the discriminator's shape prior, enhancing performance without retraining.

Contribution

It introduces a method to leverage stable adversarial discriminators at inference for improved test-time segmentation correction.

Findings

Test-time fine-tuning improves segmentation accuracy.

Stable discriminators effectively serve as shape priors.

Method is simple to implement and enhances existing models.

Abstract

Thanks to their ability to learn data distributions without requiring paired data, Generative Adversarial Networks (GANs) have become an integral part of many computer vision methods, including those developed for medical image segmentation. These methods jointly train a segmentor and an adversarial mask discriminator, which provides a data-driven shape prior. At inference, the discriminator is discarded, and only the segmentor is used to predict label maps on test images. But should we discard the discriminator? Here, we argue that the life cycle of adversarial discriminators should not end after training. On the contrary, training stable GANs produces powerful shape priors that we can use to correct segmentor mistakes at inference. To achieve this, we develop stable mask discriminators that do not overfit or catastrophically forget. At test time, we fine-tune the segmentor on each individual test instance until it satisfies the learned shape prior. Our method is simple to implement and increases model performance. Moreover, it opens new directions for re-using mask discriminators at inference. We release the code used for the experiments at https://vios-s.github.io/adversarial-test-time-training.