With the use of 3 camera modules mounted on the neckband, the system is able to cover a field of view of 360°, leaving no blindspots at all. This is coupled with a 12 metre detection distance which captures any personnel that are within range.

Equipped in the inner piece is a strong magnet which holds the outer piece in place. When in close proximity, the outer piece snaps perfectly in place, ensuring that the neckband rests perfectly onto the wearer’s shoulders. This also optimises stability while the wearer is moving around.

In order to implement a tailing detection system, a combination of human detection and person re-identification neural network models were employed. After numerous iterations, our system is able to achieve a relatively high recall of 94.7% and precision of 90.0%.

Data collected for detection purposes cannot be traced back to or identify a person. Each identity picked up by the system only saves essential information like timestamps and flag counts with features extracted using a list of non-understandable numbers.

Through the use of bluetooth (BLE), the neckband which handles the identification of suspicious persons can wirelessly communicate with the wearable belt which provides crucial information of the suspicious personnel to the wearer. This communication is immediate to ensure the safety of the wearer.

A wearable belt that upon receiving a signal, provides locational feedback of the suspicious personnel detected. This includes the direction of which the person is at together with how far the suspicious personnel is at. The feedback, provided by 8 vibration modules, is intuitive and easy to interpret.