Abstract
While self-configuring U-Net architectures excel at a vast majority of supervised medical image segmentation tasks, they strongly rely on the chosen loss function. We demonstrate that a commonly employed Dice or cross entropy loss leads to a bias of the trained network, that is critical for the clinical application of airway segmentation from CT scans. The effort to produce the most accurate segmentations is skewed towards larger anatomical structures, leaving smaller peripheral airways with poorer quality. To address this bias, we explore several different choices of amending the label definition, including morphological dilation, and find that separating the binary airway segmentations into at least two distinct structures yields substantial improvements of approximately 4% in peripheral areas. This finding could directly benefit several clinically relevant tasks, among others virtual CT bronchoscopy.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Similar content being viewed by others
References
Reichl T, Luo X, Menzel M, Hautmann H, Mori K, Navab N. Hybrid electromagnetic and image-based tracking of endoscopes with guaranteed smooth output. IJCARS. 2013;8:955– 65.
Falta F, Hansen L, Himstedt M, Heinrich MP. Learning an airway atlas from lung CT using semantic inter-patient deformable registration. Proc BVM. 2022:75–80.
Chauhan NS, Sood D, Takkar P, Dhadwal DS, Kapila R. Quantitative assessment of airway and parenchymal components of chronic obstructive pulmonary disease using thin-section helical computed tomography. Pol J Radiol. 2019;84:54–60.
Zhang M, Wu Y, Zhang H, Qin Y, Zheng H, Tang W et al. Multi-site, multi-domain airway tree modeling. Med Image Anal. 2023;90:102957.
Isensee F, Jaeger PF, Kohl SA, Petersen J, Maier-Hein KH. nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation.Nat Methods. 2021;(2):203– 11.
Tan Z, Feng J, Zhou J. SGNet: structure-aware graph-based network for airway semantic segmentation. Proc MICCAI. 2021:153–63.
Paetzold JC, Shit S, Ezhov I, Tetteh G, Ertürk A, Munich HZ et al. clDice—A novel connectivity-preserving loss function for vessel segmentation. Medical Imaging Meets NeurIPS 2019 Workshop. 2019.
Mishra D, Chaudhury S, Sarkar M, Soin AS. Ultrasound image segmentation: a deeply supervised network with attention to boundaries. IEEE Trans Biomed Eng. 2018;66(6):1637– 48.
Eisenmann M, Reinke A, Weru V, Tizabi MD, Isensee F, Adler TJ et al. Why is the winner the best? Proc IEEE CVPR. 2023:19955–66.
Armato III SG, McLennan G, Bidaut L, McNitt-Gray MF, Meyer CR, Reeves AP et al. The lung image database consortium (LIDC) and image database resource initiative (IDRI): a completed reference database of lung nodules on CT scans. Med Phys. 2011;38(2):915–31.
Wasserthal J, Breit HC, Meyer MT, Pradella M, Hinck D, Sauter AWet al. Totalsegmentator: robust segmentation of 104 anatomic structures in ct images. Radiol Artif Intell. 2023;5(5).
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2024 Der/die Autor(en), exklusiv lizenziert an Springer Fachmedien Wiesbaden GmbH, ein Teil von Springer Nature
About this paper
Cite this paper
Falta, F., Heinrich, M.P., Himstedt, M. (2024). Addressing the Bias of the Dice Coefficient. In: Maier, A., Deserno, T.M., Handels, H., Maier-Hein, K., Palm, C., Tolxdorff, T. (eds) Bildverarbeitung für die Medizin 2024. BVM 2024. Informatik aktuell. Springer Vieweg, Wiesbaden. https://doi.org/10.1007/978-3-658-44037-4_66
Download citation
DOI: https://doi.org/10.1007/978-3-658-44037-4_66
Published:
Publisher Name: Springer Vieweg, Wiesbaden
Print ISBN: 978-3-658-44036-7
Online ISBN: 978-3-658-44037-4
eBook Packages: Computer Science and Engineering (German Language)