Table 7.

Comparison of different DL methods.

Reference	Classification Techniques	Data Set	Performance Evaluation
			Sensitivity of Method (%)	Specificity of the Method (%)	Precision (%)	Accuracy (%)
[19]	Transfer learning	PH2	98.4	98.8	97.7	98.7
[20]	Alex Net TL	PH2	100	96.7	Not given	97.5
[21]	DCNN	PH2	72	89	Not given	80.5
[22]	FRCN (Full resolution Conv. Network)	PH2	91.6	96.5	Not given	94.6
[23]	HRFB (High resolution Feature blocks)	PH2	96.44	94.2	Not given	94.9
[24]	3D CTF (Color Text features)	PH2	98.2	93.8	Not given	97.5
[25]	Depth-wise residual convolutional network	PH2	100	Not given	90.1	96.5
[26]	Transfer learning	PH2	92.5	94.5	Not given	93.3
[27]	DCNN (pixel-wise)	PH2	93.1	95.1	Not given	95.4
[28]	FCNN + Google Net	ISBI (2016) challenge data set	69.1	93.6	Not given	88.2
[29]	Transfer learning	ISBI (2016) challenge data set	90.2	99.1	92.1	92.5
[30]	Fusion Method (DCNN + Features)	ISBI (2016) challenge data set	93.2	80.5	Not given	95.6
[31]	OCF (Optimized color features) + DCNN	ISBI (2016) challenge data set	92.1	90.1	Not given	92.2
[32]	Fusion method (DCNN + Feature vectors)	ISBI (2016) challenge data set	Not given	Not given	68.9	86.9
[33]	FCN + Google Net	ISBI (2017) challenge data set	81.3	86.3	Not given	85.4
[34]	LDA + CNN	ISBI (2017) challenge data set	52.5	97.6	55.3	85.4
[35]	Transfer Learning	ISBI (2017) challenge data set	95.6	95.3	97.4	95.6
[36]	DCNN + Augmentation Algorithm	ISBI (2017) challenge data set	Not given	Not given	73.9	89.2
[37]	CNN + Ranking Algorithm + Ra Pooling	ISBI (2017) challenge data set	60	88.7	Not given	84.4
[38]	Transfer Learning Algorithm	ISBI (2018) challenge data set	80.2	98.1	Not given	97.6
[39]	CNN + Regularize	ISIC data set	94.3	93.2	Not given	97.6
[40]	ECOC + SVM + DCNN	Random images data set	97.0	90.2	Not given	94.3
[41]	Fusion method (Alexnet + VGG16)	Multiple	99.3	98.4	Not given	99%
[42]	Modified CNN	Multiple Dermis + Der Quest	94.2	94.5	Not given	94.6