Skip to main content

Table 1 Comparison of parameters, FLOPs, and accuracy for tested network architectures on the ISIC dataset

From: Cancer-Net SCa: tailored deep neural network designs for detection of skin cancer from dermoscopy images

Paper

Architecture

Parameters (M)

FLOPs (G)

Accuracy (%)

Budhiman et al. [19]

ResNet-50 [38]

23.52

7.72

78.3

Demir et al. [20]

Inception V3 [18]

23.80

43.6

84.2

Hassan et al. [58]

DenseNet-121 [57]

7.00

2.80

83.9

Ech-Cherif et al. [59]

MobileNetV2 [45]

4.20

0.57

83.9

 

Cancer-Net SCa-A

13.65

4.66

83.7

 

Cancer-Net SCa-B

0.80

0.43

84.4

 

Cancer-Net SCa-C

1.19

0.40

83.9

  1. Best results highlighted in bold