MAM: Multiple Attention Mechanism Neural Networks for Cross-Age Face Recognition

Table 6

Rank-1 identification rates of the proposed MAM-CNN and other methods on MORPH dataset.

Model	Rank-1 (%)
HFA [8]	91.14
CARC [20]	92.80
MEFA [10]	93.80
FaceNet [13]	95.67
LF-CNNs [16]	97.51
AE-CNNs [15]	98.13
ResNet-50(fine-tuned by MORPH)	95.91
Residual attention ResNet50 with HRC	97.53
Residual attention ResNet50 with HRC + self-attention	98.84
MAM-CNN	99.07