Gait recognition and fall detection with inertial sensors

Abstract

In contrast to visual information that is recorded by cameras placed somewhere, inertial information can be obtained from mobile phones that are commonly used in daily life. We present in this talk a general deep learning approach for gait and soft biometrics (age and gender) recognition. Moreover, we also study the use of gait information to detect actions during walking, specifically, fall detection. We perform a thorough experimental evaluation of the proposed approach on different datasets: OU-ISIR Biometric Database, DFNAPAS, SisFall, UniMiB-SHAR and ASLH. The experimental results show that inertial information can be used for gait recognition and fall detection with state-of-the-art results.