The $7 device detects hidden cameras in 5 seconds
The technology called SweepLED is led by Professor Han Jun of the Department of Computing - Korea Advanced Institute of Science and Technology (KAIST), in cooperation with the National University of Singapore (NUS) and Singapore Management University (SMU). The research results were reported at the International Conference on Mobile Computing (ACM MobiSys 2026) in June.
Unlike traditional methods that require users to use a flashlight to find reflective points or use an infrared scanning camera, which is time-consuming and easily confused with metal, glass or glossy plastic surfaces, SweepLED solves the problem thanks to deep learning algorithms.
According to Tom's Hardware, the accessory has an LED grid design, easily attached to the back of the smartphone with a magnet. When operating, the light system continuously changes lighting direction. Light reflected from common objects will move or disappear according to the angle of projection. However, due to its complex optical structure including many layers of lenses, apertures and image sensors, the camera lens will create a very specific light reflection pattern when projected from many different angles.
The AI application installed on the phone will analyze changes in reflection patterns in real time instead of just identifying a single point of light. In practical testing on 30 familiar household items such as phone chargers, alarm clocks, remote controls and decorations, SweepLED achieved 93.9% accuracy, giving results in less than 5 seconds.
The situation of installing hidden cameras in private areas such as toilets, dressing rooms, rented houses or hotel rooms is becoming a problem in many countries. Professor Han Jun said that combining low-cost hardware with AI creates a convenient privacy protection tool, helping even those who are not tech-savvy to protect themselves.
However, experts note that this $7 tool only supports detecting hidden cameras on site. The device cannot prevent risks from tens of thousands of vulnerable security cameras streaming public data globally, or the popularity of smart glasses capable of secretly recording.
Huy Duc