Transformer for Age-Invariant Face Recognition

Table 1. Information about the used dataset
View table
Table 1. Information about the used dataset
Dataset Number of identities Number of images
faces emore^［11］ 85742 5774205
AgeDB-30^［13］ 6000 pairs 12000
CACD_VS^［14］ 4000 pairs 8000
CALFW^［15］ 6000 pairs 12000
LFW^［16］ 6000 pairs 12000

Table 2. Comparison of parameters of different models
View table
Table 2. Comparison of parameters of different models
Parameter Proposed model DAL（OE-CNNs）^［4］ MTLFace^［7］
MACs /GB 1.2 8.15 6.32
Params /10⁶ 40.71 57.08 47.25
Inference speed /ms Batch size is 1 13.75 25.01 21.99
Batch size is 16 1.19 2.69 2.54

Table 3. Accuracy of different methods on four datasets

View table

Table 3. Accuracy of different methods on four datasets

Method	AgeDB-30^［13］	CACD_VS^［14］	CALFW^［15］	LFW^［16］
DAL（OE-CNNs）^［4］		99.4		99.47
DAL（OE-CNN_s）-Re	95.98	99.38	95.48	99.56
MTLFace^［7］	96.23	99.55	95.62	99.52
Proposed method	96.25	99.35	95.48	99.63

Tools

Get Citation

Copy Citation Text

Cheng Liu, Liangcai Cao, Ye Jin, Haowei Wang, Songfeng Yin. Transformer for Age-Invariant Face Recognition[J]. Laser & Optoelectronics Progress, 2023, 60(10): 1010019

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites