The value of measuring uncertainty in neural networks in dermoscopy

J Am Acad Dermatol. 2022 Nov;87(5):1191-1193. doi: 10.1016/j.jaad.2022.03.006. Epub 2022 Mar 8.

Authors

Pieter Van Molle¹, Lieve Brochez², Tim Verbelen³, Cedric De Boom³, Bert Vankeirsbilck³, Evelien Verhaeghe², Sofie Mylle², Pieter Simoens³, Bart Dhoedt³

Affiliations

¹ Department of Information Technology, IDLab, Ghent University - imec, Gent, Belgium. Electronic address: pieter.vanmolle@ugent.be.
² Ghent University Hospital, Gent, Belgium.
³ Department of Information Technology, IDLab, Ghent University - imec, Gent, Belgium.

PMID: 35276285
DOI: 10.1016/j.jaad.2022.03.006

No abstract available

Keywords: CNN; computer vision; deep learning; melanoma; neural networks; skin lesion classification; survey; uncertainty.

Publication types

Research Support, Non-U.S. Gov't

MeSH terms

Dermoscopy
Humans
Melanoma*
Neural Networks, Computer
Skin Neoplasms*
Uncertainty