Details
| Originalsprache | Englisch |
|---|---|
| Titel des Sammelwerks | IEEE International Symposium on Biomedical Imaging |
| Untertitel | ISBI 2024 - Conference Proceedings |
| Herausgeber (Verlag) | IEEE Computer Society |
| Seitenumfang | 5 |
| ISBN (elektronisch) | 9798350313338 |
| ISBN (Print) | 979-8-3503-1334-5 |
| Publikationsstatus | Veröffentlicht - 2024 |
| Veranstaltung | 21st IEEE International Symposium on Biomedical Imaging, ISBI 2024 - Athens, Griechenland Dauer: 27 Mai 2024 → 30 Mai 2024 |
Publikationsreihe
| Name | Proceedings - International Symposium on Biomedical Imaging |
|---|---|
| ISSN (Print) | 1945-7928 |
| ISSN (elektronisch) | 1945-8452 |
Abstract
Vertebral body (VB) fractures can have severe implications for the patient's well-being, but often remain undetected. Automatic vertebral fracture assessment is challenging due to severe class imbalance in fracture grades, scarcity of annotated data, and often subtle visual differences between the grades. In this paper, we show that leveraging unlabeled data, collected from eight diverse, publicly available datasets in a self-supervised way via the BYOL framework, can noticeably improve performance. Our pre-trained models outperformed the non-pre-trained baselines in 17 out of 18 comparisons, 11 of which were statistically significant, when evaluated on the widely-used VerSe dataset, our in-house dataset, and combinations thereof. Furthermore, our models reached an average AUROC value of 99.6% on the VerSe test corpus, the highest recorded in the literature so far.
ASJC Scopus Sachgebiete
- Ingenieurwesen (insg.)
- Biomedizintechnik
- Medizin (insg.)
- Radiologie, Nuklearmedizin und Bildgebung
Zitieren
- Standard
- Harvard
- Apa
- Vancouver
- BibTex
- RIS
IEEE International Symposium on Biomedical Imaging: ISBI 2024 - Conference Proceedings. IEEE Computer Society, 2024. (Proceedings - International Symposium on Biomedical Imaging).
Publikation: Beitrag in Buch/Bericht/Sammelwerk/Konferenzband › Aufsatz in Konferenzband › Forschung › Peer-Review
}
TY - GEN
T1 - Using Unlabeled Data in Self-Supervised Training Improves Automatic Vertebral Body Fracture Assessment
AU - Laue, Julian Lukas
AU - Adolph, Christian
AU - Yilmaz, Eren Bora
AU - Meyer, Carsten
N1 - Publisher Copyright: © 2024 IEEE.
PY - 2024
Y1 - 2024
N2 - Vertebral body (VB) fractures can have severe implications for the patient's well-being, but often remain undetected. Automatic vertebral fracture assessment is challenging due to severe class imbalance in fracture grades, scarcity of annotated data, and often subtle visual differences between the grades. In this paper, we show that leveraging unlabeled data, collected from eight diverse, publicly available datasets in a self-supervised way via the BYOL framework, can noticeably improve performance. Our pre-trained models outperformed the non-pre-trained baselines in 17 out of 18 comparisons, 11 of which were statistically significant, when evaluated on the widely-used VerSe dataset, our in-house dataset, and combinations thereof. Furthermore, our models reached an average AUROC value of 99.6% on the VerSe test corpus, the highest recorded in the literature so far.
AB - Vertebral body (VB) fractures can have severe implications for the patient's well-being, but often remain undetected. Automatic vertebral fracture assessment is challenging due to severe class imbalance in fracture grades, scarcity of annotated data, and often subtle visual differences between the grades. In this paper, we show that leveraging unlabeled data, collected from eight diverse, publicly available datasets in a self-supervised way via the BYOL framework, can noticeably improve performance. Our pre-trained models outperformed the non-pre-trained baselines in 17 out of 18 comparisons, 11 of which were statistically significant, when evaluated on the widely-used VerSe dataset, our in-house dataset, and combinations thereof. Furthermore, our models reached an average AUROC value of 99.6% on the VerSe test corpus, the highest recorded in the literature so far.
KW - BYOL
KW - Deep Learning
KW - Self-Supervised Learning
KW - Vertebral Fracture Assessment
UR - http://www.scopus.com/inward/record.url?scp=85203344352&partnerID=8YFLogxK
U2 - 10.1109/ISBI56570.2024.10635829
DO - 10.1109/ISBI56570.2024.10635829
M3 - Conference contribution
AN - SCOPUS:85203344352
SN - 979-8-3503-1334-5
T3 - Proceedings - International Symposium on Biomedical Imaging
BT - IEEE International Symposium on Biomedical Imaging
PB - IEEE Computer Society
T2 - 21st IEEE International Symposium on Biomedical Imaging, ISBI 2024
Y2 - 27 May 2024 through 30 May 2024
ER -