Decomposition Sampling for Efficient Region Annotations in Active Learning

Qiu J, Wilm F, Öttl M, Utz J, Schlereth M, Schillinger M, Aubreville M, Breininger K (2026)

Publication Type: Conference contribution

Publication year: 2026

Publisher: Institute of Electrical and Electronics Engineers Inc.

Pages Range: 4160-4169

Conference Proceedings Title: Proceedings - 2026 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2026

Event location: Tucson, AZ US

ISBN: 9798331555115

DOI: 10.1109/WACV61042.2026.00405

Abstract

Active learning improves annotation efficiency by selecting the most informative samples for annotation and model training. While most prior work has focused on selecting informative images for classification tasks, we investigate the more challenging setting of dense prediction, where annotations are more costly and time-intensive, especially in medical imaging. Region-level annotation has been shown to be more efficient than image-level annotation for these tasks. However, existing methods for representative annotation region selection suffer from high computational and memory costs, irrelevant region choices, and heavy reliance on uncertainty sampling. We propose decomposition sampling (DECOMP), a new active learning sampling strategy that addresses these limitations. It enhances annotation diversity by decomposing images into class-specific components using pseudo-labels and sampling regions from each class. Class-wise predictive confidence further guides the sampling process, ensuring that difficult classes receive additional annotations. Across ROI classification, 2-D segmentation, and 3-D segmentation, DECOMP consistently surpasses baseline methods by better sampling minority-class regions and boosting performance on these challenging classes. Code is in https://github.com/JingnaQiu/DECOMP.git.

Authors with CRIS profile

Jingna Qiu Juniorprofessur für Artificial Intelligence in Medical Imaging Frauke Wilm Juniorprofessur für Artificial Intelligence in Medical Imaging Mathias Öttl Juniorprofessur für Artificial Intelligence in Medical Imaging Jonas Utz Department Artificial Intelligence in Biomedical Engineering (AIBE) Maja Schlereth Department of Medicine 3 – Rheumatology and Immunology Moritz Schillinger Department Artificial Intelligence in Biomedical Engineering (AIBE) Katharina Breininger

Additional Organisation(s)

Erlangen National High Performance Computing Center (NHR@FAU) Sonderforschungsbereich 1540/1 - Erforschung der Mechanik des Gehirns (EBM): Verständnis, Engineering und Nutzung mechanischer Eigenschaften und Signale in der Entwicklung, Physiologie und Pathologie des zentralen Nervensystems

Involved external institutions

Fachhochschule Flensburg DE Germany (DE) Julius-Maximilians-Universität Würzburg DE Germany (DE)

How to cite

APA:

Qiu, J., Wilm, F., Öttl, M., Utz, J., Schlereth, M., Schillinger, M.,... Breininger, K. (2026). Decomposition Sampling for Efficient Region Annotations in Active Learning. In Proceedings - 2026 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2026 (pp. 4160-4169). Tucson, AZ, US: Institute of Electrical and Electronics Engineers Inc..

MLA:

Qiu, Jingna, et al. "Decomposition Sampling for Efficient Region Annotations in Active Learning." Proceedings of the 2026 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2026, Tucson, AZ Institute of Electrical and Electronics Engineers Inc., 2026. 4160-4169.

BibTeX: Download