Merging public datasets and rejecting high-entropy predictions boosts reported melanoma detection accuracy to 97.8% and cuts misdiagnoses by over 40%, but lacks baseline comparisons and error bars.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Melanoma Detection with Uncertainty Quantification
Merging public datasets and rejecting high-entropy predictions boosts reported melanoma detection accuracy to 97.8% and cuts misdiagnoses by over 40%, but lacks baseline comparisons and error bars.