safety security
A Faint Image Trace Survives Model Training
SpectralCanary blends one low-frequency signal through a collection, then looks for its transfer into a fine-tuned image model's outputs.
Summary
SpectralCanary blends one low-frequency signal through a collection, then looks for its transfer into a fine-tuned image model's outputs.
Dataset-use audits usually struggle when a model absorbs a collection's style without copying individual images. SpectralCanary plants a smooth low-pass version of a fixed canary image into every released image while leaving captions untouched. A lightweight verifier then checks generated outputs for the transferred signal. Across the reported tests, per-generation detection averaged 92.4 percent at a 1 percent false-positive rate over five datasets with SD3.5, while SSIM remained 0.989. The strongest evaluated removal attack caused more than 25 FID points of degradation. The claims are limited to the architectures, transformations, datasets and attack strategies tested.
Why it matters
SpectralCanary blends one low-frequency signal through a collection, then looks for its transfer into a fine-tuned image model's outputs.
Limits and context
No additional limitation was separately recorded.
Key claims
SpectralCanary blends one low-frequency signal through a collection, then looks for its transfer into a fine-tuned image model's outputs.
Evidence: source-2026-10-23-004
Sources
- arXiv preprint 2610.31659, original v1arXiv · primary research
Corrections
No corrections have been recorded for this story.