Abstract
In its 4th edition, the Diabetic Foot Ulcer Challenge (DFUC) 2024 focuses on the task of Diabetic Foot Ulcer (DFU) segmentation by involving aspects of self-supervised learning as contrast to the DFUC 2022. This paper provides information on the challenge background, outlines details on an synthetic extension of the DFU dataset, elaborates rules for participation, and reports performance of two vastly different baseline methods that highlight the potential diversity of welcome approaches. It further provides a condensed summary on final challenge results and methods employed by participating teams. A discussion puts the different approaches and outcomes into perspective, considering their creativity and applicability on a larger scale. Eventually, a retrospective on the DFUC 2024 finds the given task to be considerably challenging, inspiring diverse contributions. Even though accurate DFU segmentation currently seems out of reach when involving a higher level of self-supervision, respective contributions demonstrate promising kick-off strategies for scenarios in which unlabelled data needs to be harnessed.