Label Uncertainty for Ultrasound Segmentation Article Swipe

MAKE IMPACT

Malini Shivaram , Gautam Rajendrakumar Gare , Laura Hutchins , Jane Duplantis , Thomas Deiss , Tânia Gomes , Thong Tran , K.H. Patel , Thomas H. Fox , Amita Krishnan , Deva Ramanan , Bennett P. deBoisblanc , Ricardo Luis Rodriguez , John Galeotti · YOU? · · 2025 · Open Access · · DOI: https://doi.org/10.48550/arxiv.2508.15635

In medical imaging, inter-observer variability among radiologists often introduces label uncertainty, particularly in modalities where visual interpretation is subjective. Lung ultrasound (LUS) is a prime example-it frequently presents a mixture of highly ambiguous regions and clearly discernible structures, making consistent annotation challenging even for experienced clinicians. In this work, we introduce a novel approach to both labeling and training AI models using expert-supplied, per-pixel confidence values. Rather than treating annotations as absolute ground truth, we design a data annotation protocol that captures the confidence that radiologists have in each labeled region, modeling the inherent aleatoric uncertainty present in real-world clinical data. We demonstrate that incorporating these confidence values during training leads to improved segmentation performance. More importantly, we show that this enhanced segmentation quality translates into better performance on downstream clinically-critical tasks-specifically, estimating S/F oxygenation ratio values, classifying S/F ratio change, and predicting 30-day patient readmission. While we empirically evaluate many methods for exposing the uncertainty to the learning model, we find that a simple approach that trains a model on binarized labels obtained with a (60%) confidence threshold works well. Importantly, high thresholds work far better than a naive approach of a 50% threshold, indicating that training on very confident pixels is far more effective. Our study systematically investigates the impact of training with varying confidence thresholds, comparing not only segmentation metrics but also downstream clinical outcomes. These results suggest that label confidence is a valuable signal that, when properly leveraged, can significantly enhance the reliability and clinical utility of AI in medical imaging.

Type

preprint

Language

en

Landing Page

http://arxiv.org/abs/2508.15635

PDF

https://arxiv.org/pdf/2508.15635

OA Status

green

OpenAlex ID

https://openalex.org/W4416051353

Raw OpenAlex JSON

OpenAlex ID

https://openalex.org/W4416051353

Canonical identifier for this work in OpenAlex

DOI

https://doi.org/10.48550/arxiv.2508.15635

Digital Object Identifier

Title

Label Uncertainty for Ultrasound Segmentation

Work title

Type

preprint

OpenAlex work type

Language

en

Primary language

Publication year

2025

Year of publication

Publication date

2025-08-21

Full publication date if available

Authors

Malini Shivaram, Gautam Rajendrakumar Gare, Laura Hutchins, Jane Duplantis, Thomas Deiss, Tânia Gomes, Thong Tran, K.H. Patel, Thomas H. Fox, Amita Krishnan, Deva Ramanan, Bennett P. deBoisblanc, Ricardo Luis Rodriguez, John Galeotti

List of authors in order

Landing page

https://arxiv.org/abs/2508.15635

Publisher landing page

PDF URL

https://arxiv.org/pdf/2508.15635

Direct link to full text PDF

Open access

Yes

Whether a free full text is available

OA status

green

Open access status per OpenAlex

OA URL

https://arxiv.org/pdf/2508.15635

Direct OA link when available

Cited by

0

Total citation count in OpenAlex

Full payload