Table 5.

Comparison of eXnet with benchmark networks on the CK+ dataset.

Model	Accuracy (%) 10-Cross
VGG [7]	94.6
ResNet [20]	94
DenseNet [47]	92
DeXpression [13]	96
Tautkute et al. [10]	92
Lopes et al. [16]	92.73
Jain et al. [17]	93.24
Shao et al. [18] light-CNN	92.86 (without 10-cross )
Shao et al. [18] pretrained-CNN	95.29 (without 10-cross)
eXnet (ours)	95.63
eXnet${}^{cutout}$ (ours)	95.81
eXnet${}^{mixup}$(ours)	96.17
eXnet${}^{cm}$ (ours)	96.75