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

Table 1

Summary of methods on cross-age face recognition problem.

Method	Generative or discriminative	Key innovations
[5]	Generative	Proposed a 3D aging modeling technique
[6]	Generative	Proposed a sparse-constrained NMF algorithm
[7]	Discriminative	SIFT and multi-scale LBP descriptor
[8]	Discriminative	Use HFA to separate the age and the identity information
[9]	Discriminative	Formulate it as a graph matching problem
[10]	Discriminative	Proposed a maximum entropy feature descriptor
[15]	Discriminative	Proposed an age estimation guided convolutional neural network
[16]	Discriminative	Proposed a latent factor guided CNN framework
[17]	Discriminative	Proposed an age-related factor guided joint task modeling CNN network