CelebA-Spoof

该数据集名为CelebA-Spoof，是最近发布的、目前公开可用的最大反欺骗任务数据集。为确保测试的公正性，该数据集在训练集和测试集中确保了面部识别数据无交叉，从而保持了测试集的完整性。该数据集的主要任务是反欺骗。

The CelebA-Spoof dataset is the largest publicly available anti-spoofing dataset released recently. To ensure test fairness, this dataset eliminates any crossover of facial recognition data between the training set and the test set, thus maintaining the integrity of the test set. The core task of this dataset is anti-spoofing.