arXiv：cs.LG· Danil Davydov, Bader Rasheed, Dmitriy Vatolin·· 4 小时前
Noiseprint++ 噪声残差用于换脸检测：冗余而非互补，原因何在
Camera-Noise Residuals for Face-Swap Detection: Redundant, Not Complementary, and Why
AI 导读
在 FaceForensics++ 上的三模型消融实验显示，将 Noiseprint++ 噪声残差通道与 RGB Xception 主干融合，效果并不优于单用 RGB，且残差分支单独使用接近随机水平。
正文
Abstract:Fusing a learned camera-noise fingerprint with an RGB appearance backbone is an appealing route to generator-independent deepfake detection, because the noise residual is grounded in image-formation physics rather than in the texture statistics of a particular generator. We test, on FaceForensics++, whether a Noiseprint++ residual channel carries information \emph{complementary} to an RGB Xception backbone for face-swap detection. A three-model ablation (RGB-only, residual-only, late-fusion) shows that fusion does not improve over RGB alone and that the residual branch alone is near chance. A seven-level bottleneck diagnostic localizes the cause: the noise maps do carry a discriminative signal, but it is statistical---carried by the per-sample first and second moments (mean, variance, energy) of the residual---and the per-sample \texttt{InstanceNorm} layer placed at the noise-branch input, following the TruFor template, standardizes exactly those moments away (five-fold cross-validated AUC drops from $0.747$ to $0.554$). A context-crop control rules out cropping geometry, and two fixed-fusion variants that remove the bottleneck recover the statistical signal yet still fail to beat RGB on every dataset. We conclude that, on this manipulation distribution, the noise residual is redundant with RGB rather than complementary, and we give concrete guidance for practitioners adopting noise-residual fusion for face-swap detection.
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.11683 [cs.LG] |
| (or arXiv:2610.11683v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.11683 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Bader Rasheed [view email]
[v1]
Thu, 8 Oct 2026 10:55:41 UTC (2,940 KB)
来源：arXiv：cs.LG · arxiv.org