DeeperForensics-1.0: A Large-Scale Dataset for Real-World Face Forgery Detection

TL;DR

DeeperForensics-1.0 is a large-scale, diverse face forgery detection dataset with high-quality fake videos, designed to advance research in real-world face manipulation detection.

Contribution

It introduces the largest face forgery dataset with extensive real-world perturbations and a new face swapping framework, providing a challenging benchmark for detection methods.

Findings

Existing detection models show varied performance on the new dataset.

High-quality fake videos are highly deceptive in human evaluations.

The dataset enables comprehensive evaluation of detection algorithms.

Abstract

We present our on-going effort of constructing a large-scale benchmark for face forgery detection. The first version of this benchmark, DeeperForensics-1.0, represents the largest face forgery detection dataset by far, with 60,000 videos constituted by a total of 17.6 million frames, 10 times larger than existing datasets of the same kind. Extensive real-world perturbations are applied to obtain a more challenging benchmark of larger scale and higher diversity. All source videos in DeeperForensics-1.0 are carefully collected, and fake videos are generated by a newly proposed end-to-end face swapping framework. The quality of generated videos outperforms those in existing datasets, validated by user studies. The benchmark features a hidden test set, which contains manipulated videos achieving high deceptive scores in human evaluations. We further contribute a comprehensive study that evaluates five representative detection baselines and make a thorough analysis of different settings.