Uncertainty-Aware Regularization for Image-to-Image Translation
This work addresses the need for reliable uncertainty quantification in medical imaging applications, though it appears incremental by building on existing methods for uncertainty estimation in deep networks.
The paper tackles the problem of improving uncertainty estimation in medical image-to-image translation by proposing Uncertainty-Aware Regularization (UAR), which integrates aleatoric uncertainty and uses simple priors to refine uncertainty maps and enhance reconstruction quality, demonstrating improved translation performance and better uncertainty estimations in noisy and ambiguous scenarios.
The importance of quantifying uncertainty in deep networks has become paramount for reliable real-world applications. In this paper, we propose a method to improve uncertainty estimation in medical Image-to-Image (I2I) translation. Our model integrates aleatoric uncertainty and employs Uncertainty-Aware Regularization (UAR) inspired by simple priors to refine uncertainty estimates and enhance reconstruction quality. We show that by leveraging simple priors on parameters, our approach captures more robust uncertainty maps, effectively refining them to indicate precisely where the network encounters difficulties, while being less affected by noise. Our experiments demonstrate that UAR not only improves translation performance, but also provides better uncertainty estimations, particularly in the presence of noise and artifacts. We validate our approach using two medical imaging datasets, showcasing its effectiveness in maintaining high confidence in familiar regions while accurately identifying areas of uncertainty in novel/ambiguous scenarios.