On the pitfalls of entropy-based uncertainty for multi-class semi-supervised segmentation
This addresses a critical bottleneck in semi-supervised learning for medical or multi-class segmentation, offering a domain-specific but impactful solution.
The paper tackles the problem of suboptimal performance in multi-class semi-supervised segmentation when using entropy-based uncertainty, showing that it leads to erroneous approximations due to inter-class overlap. The result is a substantial improvement in segmentation accuracy by replacing entropy with divergence distances, as demonstrated on a challenging dataset.
Semi-supervised learning has emerged as an appealing strategy to train deep models with limited supervision. Most prior literature under this learning paradigm resorts to dual-based architectures, typically composed of a teacher-student duple. To drive the learning of the student, many of these models leverage the aleatoric uncertainty derived from the entropy of the predictions. While this has shown to work well in a binary scenario, we demonstrate in this work that this strategy leads to suboptimal results in a multi-class context, a more realistic and challenging setting. We argue, indeed, that these approaches underperform due to the erroneous uncertainty approximations in the presence of inter-class overlap. Furthermore, we propose an alternative solution to compute the uncertainty in a multi-class setting, based on divergence distances and which account for inter-class overlap. We evaluate the proposed solution on a challenging multi-class segmentation dataset and in two well-known uncertainty-based segmentation methods. The reported results demonstrate that by simply replacing the mechanism used to compute the uncertainty, our proposed solution brings substantial improvement on tested setups.