CEDAR

该数据集名为CEDAR，包含了1,320个真实签名和1,320个伪造签名，被用于评估在签名验证模型上对抗性攻击的性能，特别是针对风格转移的攻击。该数据集的规模为真实和伪造签名各1,320个，任务是对签名验证模型进行对抗性攻击评估。

This dataset, named CEDAR, includes 1,320 genuine signatures and 1,320 forged signatures. It is designed to evaluate the performance of adversarial attacks on signature verification models, particularly style-transfer-based attacks. With an equal number of genuine and forged signatures (1,320 for each category), the core task of this dataset is to assess adversarial attacks against signature verification models.