Table 4. Performance comparison with related studies.

Reference	Classification type	Methodology	Dataset sample	Result
Nijhawan et al. (2017)	Multi-class classification: 11 nail diseases	Hybrid of convolutional neural network (CNNs)	N = 4,190	Accuracy 84.58%.
Kim et al. (2020)	Binary classification: onychomycosis	Deep learning	N = 90	Sensitivity 72.7%, specificity 72.9%, AUC 0.755
Regin et al. (2022)	Binary classification: nail color change	Ensemble of CNNs	N = 185	Accuracy of 95%
Jarallah et al. (2021)	Multi-class classification: healthy nail, nail hyperpigmentation, nail clubbing and nail fungus.	AlexNet	N = 280	Accuracy 92.5%
Jansen et al. (2022)	Binary classification: onychomycosis	U-Net	N = 664	Accuracy 86.49%
Hadiyoso & Aulia (2022)	Multi-class classification: koilonychia, Beau’s lines, and leukonychia	Transfer learning VGG-16	N = 333	Accuracy 96%
Goel & Nijhawan (2019)	Binary classification: onychomycosis	Transfer learning VGG-19	N = 100	Accuracy 98.5%
Indi & Gunge (2016)	Binary classification:	RGB analysis	N = 100	Accuracy 65%
This study	Multi-class classification: melanonychia, Beau’s lines, and nail clubbing	Transfer learning VGG16 and VGG19	N = 723	Accuracy 94% and 93%