New light-based watermarking could strengthen defence against deepfake videos
SMRTR summary
Cornell researchers developed a light-based watermarking technique that embeds hidden signatures in videos through subtle lighting fluctuations during filming. The system uses "noise-coded" lights that vary brightness imperceptibly, creating unique verification codes that reveal manipulations when analyzed. This physical approach makes deepfake detection more reliable as it embeds authentication directly in the recording environment rather than relying on digital file watermarks that could be bypassed.
SMRTR provides this summary for quick context. The original article belongs to Interesting Engineering.
Read the original article