跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2611.10249 · v1

Beyond Frame-Level Accuracy: Analyzing Video-Level Stability in Transformer-Based Deepfake Forensics

Syed Azmul HasanMowmita Pervin HeraAfroza Lona MimM M Jihad HossainHazrat AliMohammad Farhad Bulbul
cs.CVcs.AIcs.LG
首次提交 2026-11-07 05:37
版本更新 2026-11-07 05:37
雷达观测 2026-11-10 10:11
arXiv ↗PDF ↗

摘要

Deepfake videos generated by advanced generative models pose serious threats to digital security, misinformation control, and media forensics. Although numerous detection methods have been proposed, their reported performances are often based on heterogeneous preprocessing pipelines, inconsistent data splits, and model-specific training protocols, hindering fair architectural comparison. This work presents a controlled and reproducible comparative study of convolutional neural networks (ResNet-18 and EfficientNet-B0) and a transformer-based architecture (ViT-B/16) within a unified, face-centric deepfake detection framework. All models are trained using identical preprocessing, augmentation, optimization settings, and data splits on Celeb-DF v2 and FaceForensics++, and results are validated across three independent random seeds with bootstrapped confidence intervals. Experimental results show that while CNN-based models perform competitively at the frame level, their performance degrades after video-level aggregation. In contrast, ViT-B/16 demonstrates more stable video-level performance, achieving 92.8\% [95\% CI: 91.7--94.0] video-level accuracy and an AUC of 0.950 [95\% CI: 0.934--0.964] on Celeb-DF v2 (3-seed bootstrapped). Cross-dataset evaluation across the same three random seeds reveals a large, asymmetric generalization gap (AUC 0.79 FF++$\rightarrow$Celeb-DF vs. 0.64 Celeb-DF$\rightarrow$FF++), indicating that in-domain video-level stability does not transfer to unseen manipulation domains. These findings indicate that global self-attention improves robustness and video-level stability for deepfake detection within a given data domain, while highlighting the need for further work on cross-domain generalization.

作者备注：Paper submitted to international journal of intelligent systems, Wiley publishers, accepted version of open-access paper

期刊信息：International Journal of Intelligent Systems, 2026, 6801006, 13 pages, 2026

查看 arXiv 原文 ↗阅读 PDF ↗

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存，登录后共享已有结果。

版本差异

目前只观测到一个版本，无法比较变化。之后收录新版本时，会在这里对照标题、作者、分类和摘要。

观测到的版本

相邻记录可能跳过中间版本；首次发现 v2 及以上也会标为版本更新。

v1 · 2026-11-07 05:37 · 最新观测版本

Beyond Frame-Level Accuracy: Analyzing Video-Level Stability in Transformer-Based Deepfake Forensics

Deepfake videos generated by advanced generative models pose serious threats to digital security, misinformation control, and media forensics. Although numerous detection methods have been proposed, their reported performances are often based on heterogeneous preprocessing pipelines, inconsistent data splits, and model-specific training protocols, hindering fair architectural comparison. This work presents a controlled and reproducible comparative study of convolutional neural networks (ResNet-18 and EfficientNet-B0) and a transformer-based architecture (ViT-B/16) within a unified, face-centric deepfake detection framework. All models are trained using identical preprocessing, augmentation, optimization settings, and data splits on Celeb-DF v2 and FaceForensics++, and results are validated across three independent random seeds with bootstrapped confidence intervals. Experimental results show that while CNN-based models perform competitively at the frame level, their performance degrades after video-level aggregation. In contrast, ViT-B/16 demonstrates more stable video-level performance, achieving 92.8\% [95\% CI: 91.7--94.0] video-level accuracy and an AUC of 0.950 [95\% CI: 0.934--0.964] on Celeb-DF v2 (3-seed bootstrapped). Cross-dataset evaluation across the same three random seeds reveals a large, asymmetric generalization gap (AUC 0.79 FF++$\rightarrow$Celeb-DF vs. 0.64 Celeb-DF$\rightarrow$FF++), indicating that in-domain video-level stability does not transfer to unseen manipulation domains. These findings indicate that global self-attention improves robustness and video-level stability for deepfake detection within a given data domain, while highlighting the need for further work on cross-domain generalization.

查看此版本解读 →查看此版本原文 ↗