Self-supervised learning methods for label-efficient dental caries classification
- High annotation costs are a substantial bottleneck in applying deep learning architectures to clinically relevant use cases, substantiating the need for algorithms to learn from unlabeled data. In this work, we propose employing self-supervised methods. To that end, we trained with three self-supervised algorithms on a large corpus of unlabeled dental images, which contained 38K bitewing radiographs (BWRs). We then applied the learned neural network representations on tooth-level dental caries classification, for which we utilized labels extracted from electronic health records (EHRs). Finally, a holdout test-set was established, which consisted of 343 BWRs and was annotated by three dental professionals and approved by a senior dentist. This test-set was used to evaluate the fine-tuned caries classification models. Our experimental results demonstrate the obtained gains by pretraining models using self-supervised algorithms. These include improved caries classification performance (6 p.p. increase in sensitivity) and, mostHigh annotation costs are a substantial bottleneck in applying deep learning architectures to clinically relevant use cases, substantiating the need for algorithms to learn from unlabeled data. In this work, we propose employing self-supervised methods. To that end, we trained with three self-supervised algorithms on a large corpus of unlabeled dental images, which contained 38K bitewing radiographs (BWRs). We then applied the learned neural network representations on tooth-level dental caries classification, for which we utilized labels extracted from electronic health records (EHRs). Finally, a holdout test-set was established, which consisted of 343 BWRs and was annotated by three dental professionals and approved by a senior dentist. This test-set was used to evaluate the fine-tuned caries classification models. Our experimental results demonstrate the obtained gains by pretraining models using self-supervised algorithms. These include improved caries classification performance (6 p.p. increase in sensitivity) and, most importantly, improved label-efficiency. In other words, the resulting models can be fine-tuned using few labels (annotations). Our results show that using as few as 18 annotations can produce >= 45% sensitivity, which is comparable to human-level diagnostic performance. This study shows that self-supervision can provide gains in medical image analysis, particularly when obtaining labels is costly and expensive.…
Author details: | Aiham TalebORCiDGND, Csaba Rohrer, Benjamin Bergner, Guilherme De Leon, Jonas Almeida Rodrigues, Falk Schwendicke, Christoph LippertORCiDGND, Joachim KroisORCiD |
---|---|
DOI: | https://doi.org/10.3390/diagnostics12051237 |
ISSN: | 2075-4418 |
Pubmed ID: | https://pubmed.ncbi.nlm.nih.gov/35626392 |
Title of parent work (English): | Diagnostics : open access journal |
Publisher: | MDPI |
Place of publishing: | Basel |
Publication type: | Article |
Language: | English |
Date of first publication: | 2022/05/16 |
Publication year: | 2022 |
Release date: | 2024/05/31 |
Tag: | annotation; data driven approaches; dental caries classification; efficient deep learning; representation learning; self-supervised learning; unsupervised methods |
Volume: | 12 |
Issue: | 5 |
Article number: | 1237 |
Number of pages: | 15 |
Funding institution: | German Research Foundation (Deutsche Forschungsgemei nschaft-DFG); [445925495]; FederalMinistry of Education and Research-Germany; (Bundesministerium fur Bildung und Forschung-BMBF) [01IS21069A]; BMBF; [BMWi 01MK21009E] |
Organizational units: | An-Institute / Hasso-Plattner-Institut für Digital Engineering gGmbH |
DDC classification: | 0 Informatik, Informationswissenschaft, allgemeine Werke / 00 Informatik, Wissen, Systeme / 004 Datenverarbeitung; Informatik |
6 Technik, Medizin, angewandte Wissenschaften / 61 Medizin und Gesundheit / 610 Medizin und Gesundheit | |
Peer review: | Referiert |
Publishing method: | Open Access / Gold Open-Access |
DOAJ gelistet | |
License (German): | CC-BY - Namensnennung 4.0 International |