Identifying Individuals Through Egocentric Motion: A Study Using Body Worn Cameras

Body worn cameras (BWCs) have become increasingly integrated into various professional contexts during the last decade, especially in law enforcement. BWCs are useful instruments for improving security, accountability, and transparency by offering real-time, first-person perspective recordings of conversations and events. They record enormous volumes of video, which can provide important insights into how people behave and act in various situations. User identification through egocentric motion analysis offers a novel perspective in biometrics, leveraging the unique motion patterns captured by BWCs. In this study, we provide FALEBego, a novel dataset using body worn camera consisting of egocentric motion of different users. We also provide some preliminary tests on the proposed dataset. This work includes two distinct insights: (1) introduction of a dataset comprising egocentric motion of 23 subjects recorded with a chest-mounted BWC, capturing their distinct walking patterns, and (2) the dataset is utilized to present a comparative analysis of different deep learning architectures for user identification based on egocentric motion data. This study highlights the potential of egocentric motion as a biometric modality and provides insights into the effectiveness of different architectures in this emerging domain. The complete dataset is available for research purposes and can be accessed by contacting the authors directly.