Overview
- A team led by KAIST with collaborators at NUS and SMU presented and published the SweepLED prototype this week at ACM MobiSys 2026.
- The clip-on uses 15 LEDs that sweep light while a phone records, and software compares LED-on and LED-off frames to isolate reflections caused by camera lenses.
- In controlled tests on 30 objects, including 12 with embedded cameras, the system reached roughly 93.9–95.1% accuracy and flagged candidates in about five seconds per object.
- Researchers say the core hardware could be built for about $7 and that SweepLED connects to an AI app, but they have not announced any commercialization plan or broad field trials.
- If validated at scale, the tool could give travelers and people in private spaces a fast, inexpensive way to check for covert cameras, though more testing is needed to prove reliability across real-world environments and device types.