CASIA-SURF

人脸反欺骗对于防止人脸识别系统出现安全漏洞至关重要。近年来，面部反欺骗基准数据集的可用性取得了很大进展。然而，现有的人脸反欺骗基准测试的学科数量（≤170）和模态（≤2）有限，阻碍了学术界的进一步发展。

为了促进未来的人脸反欺骗研究，我们发布了一个大规模的多模态数据集，即 Chalearn CASIA-SURF，它是最大的公开可用的人脸反欺骗数据集，无论是在主题还是视觉模态方面。具体来说，它由 1, 000 个主题和 21, 000 个视频组成，每个样本有 3 种模态（即 RGB、Depth 和 IR）。

Face anti-spoofing is critical for preventing security breaches in facial recognition systems. Great progress has been made in the availability of facial anti-spoofing benchmark datasets in recent years. However, the limited number of subjects (≤170) and modalities (≤2) of existing facial anti-spoofing benchmarks have hindered further academic progress.

To facilitate future research on facial anti-spoofing, we release a large-scale multimodal dataset, Chalearn CASIA-SURF, which is the largest publicly available facial anti-spoofing dataset in terms of both subjects and visual modalities. Specifically, it consists of 1,000 subjects and 21,000 videos, with each sample containing three modalities: RGB, Depth, and IR.

搜集汇总

背景与挑战

背景概述

CASIA-SURF是一个大规模多模态人脸反欺骗数据集，包含1,000个主题和21,000个视频，每个样本提供RGB、Depth和IR三种模态，是目前最大的公开可用人脸反欺骗数据集，旨在促进人脸反欺骗研究的发展。该数据集由中国科学院大学等机构于2020年发布，适用于人脸识别和安全领域的研究。