Boosting EfficientNets Ensemble Performance via Pseudo-Labels and Synthetic Images by pix2pixHD for Infection and Ischaemia Classification in Diabetic Foot Ulcers

Diabetic foot ulcers are a common manifestation of lesions on the diabetic foot, syndrome acquired as long-term complication diabetes mellitus. Accompanying neuropathy and vascular damage promote acquisition pressure injuries tissue death due to ischaemia. Affected areas prone infections, hindering healing progress. The research at hand investigates an approach classification infection ischaemia, conducted part Foot Ulcer Challenge (DFUC) 2021. Different models EfficientNet family utilized in ensembles. An extension strategy for training data is applied, involving pseudo-labeling unlabeled images, extensive generation synthetic images via pix2pixHD cope with severe class imbalances. resulting extended dataset features $8.68$ times size baseline shows real image ratio $1:3$. Performances ensembles trained compared. Synthetic featured broad qualitative variety. Results show that well their ensemble benefit from large extension. F1-Scores rare classes receive outstanding boosts, while those either not harmed or boosted moderately. A critical discussion concretizes benefits identifies limitations, suggesting improvements. work concludes performance individual can be utilizing images. Especially notably.