Through the Lens: Benchmarking Deepfake Detectors Against Moiré-Induced Distortions
Razaib Tariq, Minji Heo, Simon S. Woo, Shahroz Tariq
TL;DR
This work demonstrates that Moiré distortions from screen-captured deepfakes substantially degrade state-of-the-art detectors, with losses up to 25.4% in accuracy and further deterioration when synthetic Moiré or compression artifacts are present. The authors introduce the DeepMoiréFake (DMF) dataset, combining real and synthetic Moiré patterns captured under diverse screen, device, and lighting conditions from multiple public deepfake datasets, and evaluate 15 detectors across authentic Moiré, synthetic Moiré, and compression scenarios. They also explore demoiréing as a mitigation strategy and show that removing Moiré patterns can inadvertently remove exploitable deepfake cues, reducing detector performance further in some cases. The DMF benchmark and the accompanying empirical analysis highlight the urgent need for detectors with robust invariances to Moiré distortions and compression, and provide a path forward for practical, real-world deepfake detection research and deployment.
Abstract
Deepfake detection remains a pressing challenge, particularly in real-world settings where smartphone-captured media from digital screens often introduces Moiré artifacts that can distort detection outcomes. This study systematically evaluates state-of-the-art (SOTA) deepfake detectors on Moiré-affected videos, an issue that has received little attention. We collected a dataset of 12,832 videos, spanning 35.64 hours, from the Celeb-DF, DFD, DFDC, UADFV, and FF++ datasets, capturing footage under diverse real-world conditions, including varying screens, smartphones, lighting setups, and camera angles. To further examine the influence of Moiré patterns on deepfake detection, we conducted additional experiments using our DeepMoiréFake, referred to as (DMF) dataset and two synthetic Moiré generation techniques. Across 15 top-performing detectors, our results show that Moiré artifacts degrade performance by as much as 25.4%, while synthetically generated Moiré patterns lead to a 21.4% drop in accuracy. Surprisingly, demoiréing methods, intended as a mitigation approach, instead worsened the problem, reducing accuracy by up to 17.2%. These findings underscore the urgent need for detection models that can robustly handle Moiré distortions alongside other realworld challenges, such as compression, sharpening, and blurring. By introducing the DMF dataset, we aim to drive future research toward closing the gap between controlled experiments and practical deepfake detection.