an Integrating System of Gait Recognition by using Metric learning

Abstract

In the field of pattern recognition, gait offers significant potential as a biometric. To date, data based on the features of a person's gait have gatherers using different sources, e.g., external sensors cameras. In contrast, recent developments in miniaturization have increased the use of internal sensors as smartphone accelerometer. Furthermore, the latest technological advances may offer the capability to engage in even more low-key data collection utilizing, e.g., The triple-axis accelerometers contained in most commercially available smartphones. Nonetheless, the use of only one data source may lead to some of the dynamic features of a person's gait may be lost, meaning that complex covariate conditions could significantly influence any such recognition system.