Weakly supervised learning with positive and unlabeled data for automatic brain tumor segmentation

• Verfasst von: Wolf, Daniel [VerfasserIn]
• Verfasst von: Regnery, Sebastian [VerfasserIn]
• Verfasst von: Tarnawski, Rafal [VerfasserIn]
• Verfasst von: Bobek-Billewicz, Barbara [VerfasserIn]
• Verfasst von: Polańska, Joanna [VerfasserIn]
• Verfasst von: Götz, Michael [VerfasserIn]
• Titel: Weakly supervised learning with positive and unlabeled data for automatic brain tumor segmentation
• Verf.angabe: Daniel Wolf, Sebastian Regnery, Rafal Tarnawski, Barbara Bobek-Billewicz, Joanna Polańska and Michael Götz
• E-Jahr: 2022
• Jahr: 24 October 2022
• Umfang: 14 S.
• Fussnoten: Gesehen am 20.01.2023
• Titel Quelle: Enthalten in: Applied Sciences
• Ort Quelle: Basel : MDPI, 2011
• Jahr Quelle: 2022
• Band/Heft Quelle: 12(2022), 21, Artikel-ID 10763, Seite 1-14
• ISSN Quelle: 2076-3417
• DOI: doi:10.3390/app122110763
• Datenträger: Online-Ressource
• Sprache: eng
• Sach-SW: image segmentation
• Sach-SW: machine learning
• Sach-SW: MRI
• Sach-SW: PU-learning
• Sach-SW: random forests
• Sach-SW: semi-supervised learning
• Sach-SW: sparse annotation
• Sach-SW: tumor segmentation
• Sach-SW: weak annotations
• Sach-SW: weak supervision
• K10plus-PPN: 1831591367

Abstract

A major obstacle to the learning-based segmentation of healthy and tumorous brain tissue is the requirement of having to create a fully labeled training dataset. Obtaining these data requires tedious and error-prone manual labeling with respect to both tumor and non-tumor areas. To mitigate this problem, we propose a new method to obtain high-quality classifiers from a dataset with only small parts of labeled tumor areas. This is achieved by using positive and unlabeled learning in conjunction with a domain adaptation technique. The proposed approach leverages the tumor volume, and we show that it can be either derived with simple measures or completely automatic with a proposed estimation method. While learning from sparse samples allows reducing the necessary annotation time from 4 h to 5 min, we show that the proposed approach further reduces the necessary annotation by roughly 50% while maintaining comparative accuracies compared to traditionally trained classifiers with this approach.