S-Adapter: Generalizing Vision Transformer for Face Anti-Spoofing with Statistical Tokens

doi:doi:10.57702/uzzdf233

S-Adapter: Generalizing Vision Transformer for Face Anti-Spoofing with Statistical Tokens

Face Anti-Spoofing (FAS) aims to detect malicious attempts to invade a face recognition system by presenting spoofed faces. State-of-the-art FAS techniques predominantly rely on deep learning models but their cross-domain generalization capabilities are often hindered by the domain shift problem, which arises due to different distributions between training and testing data.

BibTex:

@dataset{Rizhao_Cai_and_Zitong_Yu_and_Chenqi_Kong_and_Haoliang_Li_and_Changsheng_Chen_and_Yongjian_Hu_and_Alex_C_Kot_2024,
    abstract = {Face Anti-Spoofing (FAS) aims to detect malicious attempts to invade a face recognition system by presenting spoofed faces. State-of-the-art FAS techniques predominantly rely on deep learning models but their cross-domain generalization capabilities are often hindered by the domain shift problem, which arises due to different distributions between training and testing data.},
    author = {Rizhao Cai and Zitong Yu and Chenqi Kong and Haoliang Li and Changsheng Chen and Yongjian Hu and Alex C. Kot},
    doi = {10.57702/uzzdf233},
    institution = {No Organization},
    keyword = {'face anti-spoofing', 'statistical tokens', 'vision transformer'},
    month = {dec},
    publisher = {TIB},
    title = {S-Adapter: Generalizing Vision Transformer for Face Anti-Spoofing with Statistical Tokens},
    url = {https://service.tib.eu/ldmservice/dataset/s-adapter--generalizing-vision-transformer-for-face-anti-spoofing-with-statistical-tokens},
    year = {2024}
}