Face Transformer for Recognition
This work addresses face recognition for computer vision applications, but it is incremental as it adapts an existing method (Transformer) to a new domain with minor modifications.
The authors investigated whether Transformer models could be used for face recognition and compared them to CNNs, finding that Face Transformer models trained on MS-Celeb-1M achieved comparable performance to CNNs with similar parameters and MACs on benchmarks like LFW and IJB-C.
Recently there has been a growing interest in Transformer not only in NLP but also in computer vision. We wonder if transformer can be used in face recognition and whether it is better than CNNs. Therefore, we investigate the performance of Transformer models in face recognition. Considering the original Transformer may neglect the inter-patch information, we modify the patch generation process and make the tokens with sliding patches which overlaps with each others. The models are trained on CASIA-WebFace and MS-Celeb-1M databases, and evaluated on several mainstream benchmarks, including LFW, SLLFW, CALFW, CPLFW, TALFW, CFP-FP, AGEDB and IJB-C databases. We demonstrate that Face Transformer models trained on a large-scale database, MS-Celeb-1M, achieve comparable performance as CNN with similar number of parameters and MACs. To facilitate further researches, Face Transformer models and codes are available at https://github.com/zhongyy/Face-Transformer.