Smartphone learns to find hidden cameras in five seconds

Engineers at the Korea Advanced Institute of Science and Technology (KAIST) have developed SweepLED, a system that turns an ordinary smartphone into a hidden-camera detector. The device uses a specialised case equipped with LEDs and machine learning algorithms to identify spy lenses in rental apartments and hotels.

Why Conventional Methods Fall Short

Traditional ways of locating hidden cameras often prove unreliable. The usual approach involves scanning a room and looking for bright reflections from lenses. The problem is that glossy buttons, metallic furniture fittings, and other shiny surfaces reflect light almost as effectively as camera lenses. As a result, sensors regularly produce false positives, leaving users without full confidence in the room's safety.

How SweepLED Works

The new system addresses this problem with a case fitted with an array of LEDs. The user points the smartphone camera at a suspicious object and holds the device steady at three different angles. Meanwhile, the LEDs scan the surface with varying illumination, and the phone's camera captures the subtlest changes in reflected light.

A neural network analyses the series of images and identifies spatiotemporal reflection patterns. Hidden camera lenses have complex internal structures, so they respond to changes in lighting angle differently from ordinary shiny objects. It is precisely this distinction that allows the system to tell a genuine lens from an everyday reflection.

How Effective Is It

In tests on 30 objects, the system achieved 94% accuracy during manual scanning. The entire process takes less than five seconds. The developers emphasise that this approach significantly reduces false alarms compared with conventional detectors.

Why This Matters

The issue of hidden cameras in rental accommodation has reached global proportions. Travellers regularly face the risk of illegal recording, and short-term rental platforms cannot always guarantee safety. Cameras are concealed in clocks, power sockets, smoke detectors, and other household items.

SweepLED allows guests to check a room using a familiar smartphone and an additional case. This makes inspection simpler and more accessible for ordinary travellers.

In Brief

Engineers at KAIST have created SweepLED—a case with LEDs and a machine learning algorithm that turns a smartphone into a hidden-camera detector. The device scans a surface from different angles and, based on reflection patterns, distinguishes camera lenses from ordinary shiny objects. In tests, accuracy reached 94%, and a scan takes less than five seconds. The development is intended to help travellers find spy cameras in rented accommodation more quickly.