Conformal Semantic Image Segmentation: Post-hoc Quantification of Predictive Uncertainty
This addresses uncertainty quantification for users of semantic image segmentation models, though it is incremental as it applies existing conformal prediction techniques to this domain.
The paper tackles the problem of quantifying predictive uncertainty in semantic image segmentation by proposing a post-hoc method using conformal prediction to generate statistically valid prediction sets guaranteed to include ground-truth masks at a predefined confidence level, demonstrating effectiveness on benchmark datasets and models.
We propose a post-hoc, computationally lightweight method to quantify predictive uncertainty in semantic image segmentation. Our approach uses conformal prediction to generate statistically valid prediction sets that are guaranteed to include the ground-truth segmentation mask at a predefined confidence level. We introduce a novel visualization technique of conformalized predictions based on heatmaps, and provide metrics to assess their empirical validity. We demonstrate the effectiveness of our approach on well-known benchmark datasets and image segmentation prediction models, and conclude with practical insights.