Impact of Benign Modifications on Discriminative Performance of Deepfake Detectors

TL;DR

This paper systematically evaluates how benign image and video processing operations affect the accuracy of deepfake detectors, highlighting the need for more robust methods in realistic scenarios.

Contribution

It introduces a rigorous framework to assess the robustness of deepfake detectors against common benign modifications, providing insights for designing more resilient detection methods.

Findings

Benign operations can significantly reduce detector accuracy

The framework quantifies the impact of each processing technique

Results highlight the need for robustness in real-world applications

Abstract

Deepfakes are becoming increasingly popular in both good faith applications such as in entertainment and maliciously intended manipulations such as in image and video forgery. Primarily motivated by the latter, a large number of deepfake detectors have been proposed recently in order to identify such content. While the performance of such detectors still need further improvements, they are often assessed in simple if not trivial scenarios. In particular, the impact of benign processing operations such as transcoding, denoising, resizing and enhancement are not sufficiently studied. This paper proposes a more rigorous and systematic framework to assess the performance of deepfake detectors in more realistic situations. It quantitatively measures how and to which extent each benign processing approach impacts a state-of-the-art deepfake detection method. By illustrating it in a popular deepfake detector, our benchmark proposes a framework to assess robustness of detectors and provides valuable insights to design more efficient deepfake detectors.