Hi there! 👋
Great work on this deepfake detection project — "towards deepfake detection that actually works" is a powerful mission.
I wanted to introduce SiliconSignature (repo | demo), an open-source image authentication system using ASIC proof-of-work + Reed-Solomon LSB watermarking.
Why this matters for deepfake detection:
Most detection is reactive — finding fakes after creation. SiliconSignature is proactive: it embeds hardware-bound watermarks into authentic media at the source. Your detector could then flag any content without a valid watermark as suspicious.
What we bring:
- 🔏 ASIC-bound proof-of-work ($10K+ to forge)
- 🛡️ Reed-Solomon survives 40% destruction, recompression, cropping
- 💻 Pure Python — integrates into any pipeline
- 📊 MIT license, zero dependencies
Collaboration angle:
Could we benchmark your detector against watermarked vs. non-watermarked forgeries? A joint paper on "reactive + proactive deepfake defense" could be very compelling.
What do you think?
Best,
Francisco Angulo de Lafuente (@Agnuxo1) — P2PCLAW
Hi there! 👋
Great work on this deepfake detection project — "towards deepfake detection that actually works" is a powerful mission.
I wanted to introduce SiliconSignature (repo | demo), an open-source image authentication system using ASIC proof-of-work + Reed-Solomon LSB watermarking.
Why this matters for deepfake detection:
Most detection is reactive — finding fakes after creation. SiliconSignature is proactive: it embeds hardware-bound watermarks into authentic media at the source. Your detector could then flag any content without a valid watermark as suspicious.
What we bring:
Collaboration angle:
Could we benchmark your detector against watermarked vs. non-watermarked forgeries? A joint paper on "reactive + proactive deepfake defense" could be very compelling.
What do you think?
Best,
Francisco Angulo de Lafuente (@Agnuxo1) — P2PCLAW