不需训练即可检测AI生成视频，靠时间特征差异识破伪造
D3: Training-Free AI-Generated Video Detection Using Second-Order Features
- 利用牛顿力学分析视频时间动态，提取二阶差分特征
- 在4个数据集上平均准确率比现有方法高10.39%
- 无需训练、计算快，适合快速部署与大规模检测
视频生成技术（如Sora）的发展使得高质量合成视频的制作变得极为便捷，引发公众对虚假内容传播的担忧。然而，现有检测方法仍受限于对合成视频中时间伪影探索不足。为此，我们基于牛顿力学建立二阶动力学分析理论框架，提出专用于时间伪影检测的二阶中心差分特征。在此基础上，我们揭示了真实视频与AI生成视频在二阶特征分布上的根本差异。具体而言，提出无需训练的检测方法D3，利用上述时间特征差异进行识别。我们在4个开源数据集（Gen-Video、VideoPhy、EvalCrafter、VidProM）共40个子集上验证其有效性。例如，在GenVideo上，D3相比之前最佳方法平均精度提升10.39%（绝对值）。额外实验表明，D3具有极高的计算效率和强鲁棒性。代码已公开于https://github.com/Zig-HS/D3。
原文摘要 · Abstract (English)
The evolution of video generation techniques, such as Sora, has made it increasingly easy to produce high-fidelity AI-generated videos, raising public concern over the dissemination of synthetic content. However, existing detection methodologies remain limited by their insufficient exploration of temporal artifacts in synthetic videos. To bridge this gap, we establish a theoretical framework through second-order dynamical analysis under Newtonian mechanics, subsequently extending the Second-order Central Difference features tailored for temporal artifact detection. Building on this theoretical foundation, we reveal a fundamental divergence in second-order feature distributions between real and AI-generated videos. Concretely, we propose Detection by Difference of Differences (D3), a novel training-free detection method that leverages the above second-order temporal discrepancies. We validate the superiority of our D3 on 4 open-source datasets (Gen-Video, VideoPhy, EvalCrafter, VidProM), 40 subsets in total. For example, on GenVideo, D3 outperforms the previous best method by 10.39% (absolute) mean Average Precision. Additional experiments on time cost and post-processing operations demonstrate D3's exceptional computational efficiency and strong robust performance. Our code is available at https://github.com/Zig-HS/D3.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。