Laplacian Segmentation Networks Improve Epistemic Uncertainty Quantification
This work addresses a critical issue for medical imaging applications where OOD detection is essential due to variations in equipment and data, representing an incremental improvement with a novel method for a known bottleneck.
The paper tackles the problem of overconfident neural networks in image segmentation, especially for out-of-distribution (OOD) images in medical domains, by proposing Laplacian Segmentation Networks (LSN) that jointly model epistemic and aleatoric uncertainty, resulting in superior OOD detection demonstrated on three datasets.
Image segmentation relies heavily on neural networks which are known to be overconfident, especially when making predictions on out-of-distribution (OOD) images. This is a common scenario in the medical domain due to variations in equipment, acquisition sites, or image corruptions. This work addresses the challenge of OOD detection by proposing Laplacian Segmentation Networks (LSN): methods which jointly model epistemic (model) and aleatoric (data) uncertainty for OOD detection. In doing so, we propose the first Laplace approximation of the weight posterior that scales to large neural networks with skip connections that have high-dimensional outputs. We demonstrate on three datasets that the LSN-modeled parameter distributions, in combination with suitable uncertainty measures, gives superior OOD detection.