Impact of Video Processing Operations in Deepfake Detection

View on arXiv ← Back to list

Authors: Yuhang Lu, Touradj Ebrahimi

Published: 2023-03-30 09:24:17+00:00

AI Summary

This paper proposes a systematic methodology for evaluating the robustness of deepfake detectors against various real-world video processing operations. It analyzes the impact of these operations on three popular deepfake detectors, providing insights for future research.

Abstract

The detection of digital face manipulation in video has attracted extensive attention due to the increased risk to public trust. To counteract the malicious usage of such techniques, deep learning-based deepfake detection methods have been developed and have shown impressive results. However, the performance of these detectors is often evaluated using benchmarks that hardly reflect real-world situations. For example, the impact of various video processing operations on detection accuracy has not been systematically assessed. To address this gap, this paper first analyzes numerous real-world influencing factors and typical video processing operations. Then, a more systematic assessment methodology is proposed, which allows for a quantitative evaluation of a detector's robustness under the influence of different processing operations. Moreover, substantial experiments have been carried out on three popular deepfake detectors, which give detailed analyses on the impact of each operation and bring insights to foster future research.

Key findings

Different deepfake detection methods exhibit varying robustness to different video processing operations. XceptionNet and CapsuleNet are heavily influenced by training data quality, while SBIs shows more resilience but still struggles with heavy compression, low resolution, and noise. Vertical flipping significantly impacts XceptionNet and CapsuleNet but not SBIs.

Approach

The authors propose a systematic evaluation framework to quantitatively assess the robustness of deepfake detectors. This involves applying various video processing operations (compression, filtering, brightness adjustments, etc.) to test datasets and measuring the detectors' performance on the modified videos.

Datasets

FaceForensics++ (FFpp)

Model(s)

XceptionNet, CapsuleNet, SBIs

Author countries

Switzerland

← Previous