ML Security PapersExposing DeepFakes via Hyperspectral Domain Mapping
Aditya Mehta , Swarnim Chaudhary , Pratik Narang , Jagat Sesh Challa
Published on arXiv
2511.11732
Output Integrity Attack
OWASP ML Top 10 — ML09
Key Finding
HSI-Detect achieves consistent improvements over RGB-only deepfake detection baselines by exposing spectral artifacts across 31 hyperspectral bands that are averaged out in standard three-channel RGB representations.
HSI-Detect
Novel technique introduced
Modern generative and diffusion models produce highly realistic images that can mislead human perception and even sophisticated automated detection systems. Most detection methods operate in RGB space and thus analyze only three spectral channels. We propose HSI-Detect, a two-stage pipeline that reconstructs a 31-channel hyperspectral image from a standard RGB input and performs detection in the hyperspectral domain. Expanding the input representation into denser spectral bands amplifies manipulation artifacts that are often weak or invisible in the RGB domain, particularly in specific frequency bands. We evaluate HSI-Detect across FaceForensics++ dataset and show the consistent improvements over RGB-only baselines, illustrating the promise of spectral-domain mapping for Deepfake detection.
Key Contributions
- HSI-Detect: a two-stage pipeline that reconstructs a 31-channel hyperspectral image from RGB input using MST++ transformer-based spectral reconstruction, then performs deepfake classification in the hyperspectral domain
- Demonstrates that hyperspectral expansion amplifies subtle manipulation artifacts introduced by generative models that are invisible or weak in the RGB domain
- Shows consistent detection improvements over RGB-only baselines on FaceForensics++
🛡️ Threat Analysis
Proposes a novel AI-generated content detection method (deepfake detection) using hyperspectral domain mapping — directly addresses output integrity by detecting whether images are authentic or GAN/diffusion-model-generated.