# On uncertainty estimation in active learning for image segmentation

arXiv: https://arxiv.org/abs/2007.06364 · HTML: https://mlsanity.com/paper/2007.06364/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 13 Jul 2020.

## Built on by (most cited first, 9 of 9)

- [Semantic Segmentation with Active Semi-Supervised Learning](https://mlsanity.com/paper/2203.10730/) (2022)
- [Multi-Task Consistency for Active Learning](https://mlsanity.com/paper/2306.12398/) (2023)
- [A Review of Bayesian Uncertainty Quantification in Deep Probabilistic Image Segmentation](https://mlsanity.com/paper/2411.16370/) (2024)
- [On the use of uncertainty in classifying Aedes Albopictus mosquitoes](https://mlsanity.com/paper/2110.15912/) (2021)
- [Semantic Segmentation with Active Semi-Supervised Representation Learning](https://mlsanity.com/paper/2210.08403/) (2022)
- [Uncertainty Driven Active Learning for Image Segmentation in Underwater Inspection](https://mlsanity.com/paper/2403.14002/) (2024)
- [Deep few-shot learning for bi-temporal building change detection](https://mlsanity.com/paper/2108.11262/) (2021)
- [Mining of Single-Class by Active Learning for Semantic Segmentation](https://mlsanity.com/paper/2307.09109/) (2023)
- [Cost-efficient Active Learning for Referring Image Segmentation and Grounding](https://mlsanity.com/paper/2608.30621/) (2026)

## Abstract (arXiv)

Uncertainty estimation is important for interpreting the trustworthiness of machine learning models in many applications. This is especially critical in the data-driven active learning setting where the goal is to achieve a certain accuracy with minimum labeling effort. In such settings, the model learns to select the most informative unlabeled samples for annotation based on its estimated uncertainty. The highly uncertain predictions are assumed to be more informative for improving model performance. In this paper, we explore uncertainty calibration within an active learning framework for medical image segmentation, an area where labels often are scarce. Various uncertainty estimation methods and acquisition strategies (regions and full images) are investigated. We observe that selecting regions to annotate instead of full images leads to more well-calibrated models. Additionally, we experimentally show that annotating regions can cut 50% of pixels that need to be labeled by humans compared to annotating full images.