Gait-Based Biometrics System
Abstract
Behavioral biometrics is the field of study related to person identification based on the way an activity is performed. Despite the difficulties of implementation and achieving high recognition metrics, this field has advantages that attract the interest of the scientific community. In the case of gait analysis, active interaction between the user and the acquisition device is not required. This enables signals to be safely assessed remotely, that is important in the times of pandemic. Furthermore, it is not easy to deliberately mimic a person’s gait. The work concerns on the development of system that enables identifying individuals based on gait with the use of wearable sensors such as accelerometers or gyroscopes. The work describes the data preprocessing pipeline and the innovative data augmentation mechanism performed with the use of generative models. The validation of the system is carried out using three different datasets collected under laboratory, semi-laboratory and field conditions. This article focuses on presenting a comprehensive solution, with a special authors’ aspect of data augmentation.
Cite
Text
Sawicki and Saeed. "Gait-Based Biometrics System." European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, 2023. doi:10.1007/978-3-031-43430-3_29Markdown
[Sawicki and Saeed. "Gait-Based Biometrics System." European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, 2023.](https://mlanthology.org/ecmlpkdd/2023/sawicki2023ecmlpkdd-gaitbased/) doi:10.1007/978-3-031-43430-3_29BibTeX
@inproceedings{sawicki2023ecmlpkdd-gaitbased,
title = {{Gait-Based Biometrics System}},
author = {Sawicki, Aleksander and Saeed, Khalid},
booktitle = {European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases},
year = {2023},
pages = {350-355},
doi = {10.1007/978-3-031-43430-3_29},
url = {https://mlanthology.org/ecmlpkdd/2023/sawicki2023ecmlpkdd-gaitbased/}
}