Gait Recognition: the Wearable Solution

Two main factors encourage new investigations regarding biometric gait recognition. First, wearable sensors allow a new approach to this problem, which does not suffer from the hindering factors affecting computer vision methods. Occlusions, camera field of view/angle, or illumination are not issues anymore, and it is possible to better focus on gait intrinsic features. Second, wearable sensors are nowadays commonly embedded in widespread mobile devices, especially smartphones. This allows setting up a gait recognition system without special equipment (either cameras or equipped floors). However, even this new recognition approach suffers from specific limitations. Ground slope, shoe heels, walking speed, can cause signal distortions. Their possible effects must be investigated and addressed. The aim of this chapter is to provide the basics to approach gait recognition by mobile wearable sensors, and sketches the most promising techniques, while listing the (few) datasets available at present to test new algorithms.