SHIELD
收藏arXiv2024-02-07 更新2024-08-06 收录
下载链接:
http://arxiv.org/abs/2402.04178v1
下载链接
链接失效反馈官方服务:
资源简介:
用于评估多模态大语言模型在人脸欺骗和伪造检测能力的基准,设计了真/假和多选题来评估这些面部安全任务中的多模态面部数据。
This benchmark is dedicated to evaluating the capabilities of multimodal large language models (MLLMs) in face spoofing and forgery detection. True/false and multiple-choice questions are designed to assess multimodal facial data across these facial security tasks.
创建时间:
2024-02-07



