Cornell Researchers Develop Invisible Light-Based Watermark To Detect Deepfakes

Cornell University researchers have created an “invisible” light-based watermarking system that embeds unique codes into the light illuminating a scene, allowing cameras to capture authentication data without special hardware. This method, called noise-coded illumination, helps detect deepfakes by comparing coded light patterns against video footage, revealing manipulations like missing sequences or inserted objects.

Want More Context? 🔎