Mind the Gap: A survey of chronic wound datasets, methods, and clinical translation.
Authors
Affiliations (4)
Affiliations (4)
- School of Computing and Mathematics, Manchester Metropolitan University, Dalton Building, Chester Street, Manchester M1 5GD, UK.
- School of Computing and Mathematics, Manchester Metropolitan University, Dalton Building, Chester Street, Manchester M1 5GD, UK. Electronic address: [email protected].
- Medical School, Faculty of Health and Medicine, Lancaster University, UK.
- Aire Innovate Ltd, UK.
Abstract
Chronic wounds, including diabetic foot ulcers, venous ulcers, arterial ulcers and pressure ulcers, represent a growing global health challenge with significant morbidity, mortality, and economic burden. Advancements in medical image analysis have shown promise in automating wound detection, segmentation, and measurement. However, translation of these research advances into clinical practice through mobile applications remains limited. This survey presents systematic analysis of the relationship between chronic wound datasets, deep learning methods, and mobile wound assessment applications. We analyse 20 datasets (14 public/partially public and 6 private), including iterative challenge datasets that extend image pool through synthetic augmentation. We review 50 deep learning methods spanning encoder-decoder architectures to self-supervised learning approaches, and evaluate 13 mobile applications (commercially deployed clinical tools and research prototypes). We examine the alignment between publicly available chronic wound datasets and the documented capabilities of mobile wound assessment mobile applications. Our analysis reveals critical misalignments between datasets, algorithms and mobile applications. Existing datasets are typically designed for algorithm benchmarking rather than clinical deployment, and therefore lack standardised acquisition protocols, scale calibration markers, depth information, detailed tissue composition annotations, and representation of diverse skin types. Conversely, reviewed mobile applications report capabilities, including automatic tissue classification, calibrated wound measurement and depth estimation, for which publicly available data is insufficient. We map documented dataset characteristics to capabilities reported among reviewed mobile applications, identify unresolved methodological challenges, and outline key priorities for the field. Based on these insights, we propose actionable recommendations for developing clinically translatable computer vision system for wound assessment. This survey is intended to serve as a reference for researchers and practitioners seeking to bridge the gap between medical imaging methodologies and practical demands of clinical wound care.