Data Augmentation for Images of Chronic Foot Wounds

Training data for Neural Networks is often scarce in the medical domain, which often results in models that struggle to generalize and consequently showpoor performance on unseen datasets. Generally, adding augmentation methods to the training pipeline considerably enhances a model’s performance. Using the dataset of the Foot Ulcer Segmentation Challenge, we analyze two additional augmentation methods in the domain of chronic foot wounds - local warping of wound edges along with projection and blurring of shapes inside wounds. Our experiments show that improvements in the Dice similarity coefficient and Normalized Surface Distance metrics depend on a sensible selection of those augmentation methods.

Metadaten
Author:	Max Gutbrod, Benedikt Geisler, David Rauber, Christoph Palm ORCiD GND
DOI:	https://doi.org/10.1007/978-3-658-44037-4_71
Parent Title (German):	Bildverarbeitung für die Medizin 2024: Proceedings, German Workshop on Medical Image Computing, March 10-12, 2024, Erlangen
Publisher:	Springer
Place of publication:	Wiesbaden
Editor:	Andreas Maier, Thomas M. Deserno, Heinz Handels, Klaus Maier-Hein, Christoph Palm, Thomas Tolxdorff
Document Type:	conference proceeding (article)
Language:	English
Year of first Publication:	2024
Release Date:	2024/03/22
First Page:	261
Last Page:	266
Institutes:	Fakultät Informatik und Mathematik
	Fakultät Informatik und Mathematik / Regensburg Medical Image Computing (ReMIC)
Begutachtungsstatus:	peer-reviewed
research focus:	Lebenswissenschaften und Ethik
Licence (German):	Keine Lizenz - Es gilt das deutsche Urheberrecht: § 53 UrhG

Open Access