The operation of the currently available methods of 'watchlist identification' at border-crossings requires officers' subjective judgement and high degree of cooperation from the people being security-checked. The use of facial masks and makeup has also posed challenges in face recognition. Hence, the screening process at border-crossings is often inaccurate and highly time-consuming, especially when a large number of subjects/suspects are to be examined. In view of these, Dr Ajay Kumar and his research team at the Department of Computing have developed Automated Watchlist Identification System that improves recognition accuracy and efficiency; thus enhancing the security checks.
Adopted multispectral biometric technology, the system generates a matching score of a subject against a pre-set watchlist based on high-resolution moving images of a subject's irises, periocular features and face in both near-infrared and visible wavelengths. Also, the subjects' body heat can be detected by using a near-infrared camera; therefore, intricate mask, heavy make-up and spoof contact lenses are no longer the tricks for suspects to fool the security check.
Apart from detecting, the system with web-based interface is user-friendly allowing users to update watchlist and identify information easily. It also generates visual and/or audio alerts whenever any watchlist image is matched or any spoof biometric sample is suspected; suspects can thus be effectively identified.