Researchers at Cornell University have developed a light-based technique that could help investigators determine whether a video has been manipulated. Rather than placing a watermark inside a video file or relying on a particular camera, the method encodes a subtle signal in the illumination of the scene being recorded.
The approach, called Noise-Coded Illumination (NCI), uses carefully controlled, noise-like changes in the brightness of light sources. These fluctuations are designed to be difficult for people to notice, but are captured by cameras recording the scene. A forensic analyst with access to the corresponding illumination codes can then examine the footage for inconsistencies that may indicate editing or synthetic additions.
Embedding a reference in the lighting
Each coded light source carries a distinct pattern. The signal does not simply attach an identifying message to the recording. Instead, it provides information that can be used to reconstruct an image of the scene as it would appear under that coded light. This gives investigators a reference against which the recorded footage can be assessed.
When content is inserted, removed or altered, the modified region may no longer contain the expected illumination pattern. For example, an object added digitally may not show the light-code characteristics present in the surrounding physical scene. Similarly, edits that remove or rearrange portions of a recording can disrupt the temporal pattern and expose discontinuities.
Designed for different kinds of manipulation
The researchers evaluated NCI against a range of spatial and temporal edits. These included compositing, deepfake content, cuts, changes to playback speed and acceleration. The study also examined how the technique performed under different conditions, including subject and camera movement, video compression, indoor and outdoor environments, and variations in signal strength.
The team demonstrated the use of as many as three separate illumination codes in a single scene. Using multiple coded light sources can make it more difficult to construct a convincing manipulated recording, because the altered footage would need to remain consistent with the information carried by each code.
A different approach to video authentication
Conventional watermarking methods generally place identifying information in the digital content or depend on the recording device or software. NCI takes a different route by placing the verification signal in the physical environment. This could be useful in settings where lighting can be controlled but the cameras recording an event cannot, such as public appearances, interviews and press conferences.
Potential users could include news organisations, fact-checkers, public institutions and security teams that need to assess the integrity of video evidence. If adopted in suitable locations, coded lighting could provide an additional source of evidence when footage is questioned.
Promising, but not a universal deepfake detector
The method has important deployment limits. It is intended for footage recorded under specially coded illumination, so it cannot be applied retrospectively to arbitrary videos that were captured under ordinary lighting. Its usefulness also depends on access to the relevant codes and on careful interpretation of the forensic evidence. The researchers describe the work as an initial step in a broader area of video verification, rather than a complete answer to synthetic media.
As generative tools make convincing alterations easier to produce, authentication methods that begin at the point of recording may become an important complement to software-based detection. Noise-Coded Illumination explores that possibility by making the physical lighting of a scene part of the evidence used to verify its video record.
Research details
The study, “Noise-Coded Illumination for Forensic and Photometric Video Analysis”, was authored by Peter Michael, Zekun Hao, Serge Belongie and Abe Davis. It was published in ACM Transactions on Graphics on 27 June 2025 and presented at SIGGRAPH 2025.
Also read: Viksit Workforce for a Viksit Bharat
Do Follow: The Mainstream LinkedIn | The Mainstream Facebook | The Mainstream Youtube | The Mainstream Twitter
About us:
The Mainstream is a premier platform delivering the latest updates and informed perspectives across the technology business and cyber landscape. Built on research-driven, thought leadership and original intellectual property, The Mainstream also curates summits & conferences that convene decision makers to explore how technology reshapes industries and leadership. With a growing presence in India and globally across the Middle East, Africa, ASEAN, the USA, the UK and Australia, The Mainstream carries a vision to bring the latest happenings and insights to 8.2 billion people and to place technology at the centre of conversation for leaders navigating the future.