Robustness of a convolutional neural network trained on dermoscopic images and challenged with close‐up images
Our findings highlight the diagnostic limitations of applying DL-CNNs trained on dermoscopic images to close-up images, with reduced sensitivity but increased specificity. The results emphasize the need to improve cross-domain adaptability. Fine-tuning of DL-CNNs on clinical images may enhance real-world diagnostic accuracy.
Authors
- Albert Rosenberger (ORCID: https://orcid.org/0000-0001-7848-1332)
- Katharina Susanne Kommoss (ORCID: https://orcid.org/0000-0003-2428-2658)
- Holger Andreas Haenssle (ORCID: https://orcid.org/0000-0001-7255-3104)
- Alexander Enk (ORCID: https://orcid.org/0000-0002-7358-185X)
- Anastasia Sophie Vollmer (ORCID: https://orcid.org/0000-0003-4200-9647)
- Julia Katharina Winkler (ORCID: https://orcid.org/0000-0003-3197-9867)
- Wilhelm Stolz (ORCID: https://orcid.org/0000-0002-4515-9395)
Institutions
- Heidelberg University (DE)
- University Hospital Heidelberg (DE)
- München Klinik (DE)
- University of Göttingen (DE)
- Heidelberg University (US)
Publication Details
- Journal
- JDDG Journal der Deutschen Dermatologischen Gesellschaft
- Published
- 2025-10-11
- DOI
- https://doi.org/10.1111/ddg.15900
- Primary Topic
- Cutaneous Melanoma Detection and Management
- Type
- article
- Field-Weighted Citation Impact
- 0.00