Table 4.

Performance analysis of the CNN-Bayesian optimization approach, Darknet-19, and Inception-v3 on the 264 testing images extracted from the ear imagery database

	Accuracy %	Sensitivity%	Specificity%	Positive predictive value (PPV) %
CNN-Bayesian optimization	98.10	98.11	99.36	98.10
Darknet-19 [39]	34.09	34.09	78.03	31.96
Inception-v3 [40]	87.50	87.50	95.83	87.76
SVM [16]	93.90	87.80	95.90	87.70