Closer Look at the Uncertainty Estimation in Semantic Segmentation under Distributional Shift

TL;DR

This paper evaluates uncertainty estimation in semantic segmentation under distributional shifts, demonstrating that simple color transformations and model ensembles significantly improve robustness and calibration, especially in cross-dataset and simulation-to-real adaptation scenarios.

Contribution

It introduces effective baseline methods using color transformations and ensembles to enhance uncertainty estimation and domain adaptation in semantic segmentation.

Findings

Color transformations serve as strong baselines for domain shift.

Ensembles improve accuracy and uncertainty calibration.

Ensemble-based self-training enhances pseudo-label quality.

Abstract

While recent computer vision algorithms achieve impressive performance on many benchmarks, they lack robustness - presented with an image from a different distribution, (e.g. weather or lighting conditions not considered during training), they may produce an erroneous prediction. Therefore, it is desired that such a model will be able to reliably predict its confidence measure. In this work, uncertainty estimation for the task of semantic segmentation is evaluated under a varying level of domain shift: in a cross-dataset setting and when adapting a model trained on data from the simulation. It was shown that simple color transformations already provide a strong baseline, comparable to using more sophisticated style-transfer data augmentation. Further, by constructing an ensemble consisting of models using different backbones and/or augmentation methods, it was possible to improve significantly model performance in terms of overall accuracy and uncertainty estimation under the domain shift setting. The Expected Calibration Error (ECE) on challenging GTA to Cityscapes adaptation was reduced from 4.05 to the competitive value of 1.1. Further, an ensemble of models was utilized in the self-training setting to improve the pseudo-labels generation, which resulted in a significant gain in the final model accuracy, compared to the standard fine-tuning (without ensemble).