Researchers reveal smartphone-based biometric gait recognition technology

Jan 30, 2019, 1:38 pm EST | Chris Burt

Researchers from the University of Pretoria and City University of Hong Kong have proposed a continuous biometric gait recognition technique based on information gathered by the accelerometers embedded in smartphones.

“Continuous User Authentication in Smartphones Using Gait Analysis” was presented at the 44th annual IEEE Industrial Electronics Society conference, and describes a system that uses native smartphone hardware with a dynamic time warping (DNW) classification algorithm and a feed forward neural network (FFNN) as an evaluation system to authenticate the user. Within a minute of detecting a biometric non-match, the system notifies the user by email of the unauthorized access, along with a time stamp of the device’s last known location.

“If the authentication outcome is positive, the authentication process continues uninterrupted in the background,” the researchers write in the paper. “If the authentication fails, the device’s location information should be sent to a predetermined email address to notify the authorized user of the device’s whereabouts.”

The method scored a sensitivity of 0.74 and specificity of 0.78 in testing, according to Tech Xplore, and the researchers acknowledge that the technology, while promising, needs further development to be practically viable.

Chinese computer vision startup Watrix AI recently raised $14 million in an early investment round to bring its image-based gait analysis to market. Other gait recognition technologies have been developed based on data collected by a pressure pad the subject walks on.