Scholarlink

Gait-based authentication is a non-invasive and cost-effective biometric modality, making it an attractive choice for enhancing user experience and security. Gait patterns become more pronounced with age due to weakness in muscular strength or the nervous system, making it challenging to recognize individuals who exhibit gait inconsistencies. Given the limited emphasis on enhancing gait identification for individuals exhibiting inconsistent gait patterns, this work introduces an enhanced identification technique to mitigate the impact of intra-subject gait fluctuation. The innovation aims to improve the system’s resilience by combining gait data from multiple inertial sensors and capturing spatial and temporal features. Experimental results showcase a high identification accuracy of \(97.22\%\) in the presence of pronounced gait patterns. Acknowledging the vulnerability to illicit access to gait readings and the underexplored domain of gait-based authentication security, this work introduces an authentication protocol called Gait4Auth. Gait4Auth utilizes a unique marker called a Temporary Authentication Code (TAC), which provides an additional validation layer that cross-checks a user’s identity. It adds complexity to unauthorized attempts at authentication thanks to its distributed nature. We present a lightweight liveness detection method by deploying decoy TACs that detect attempts at gaining unauthorized access.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Gait4Auth: Enhancing Identification and Security in Gait-Based Authentication

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Youssef Yamout,
  • Shahrear Iqbal,
  • Nilesh Chakraborty,
  • Mohammad Zulkernine

摘要

Gait-based authentication is a non-invasive and cost-effective biometric modality, making it an attractive choice for enhancing user experience and security. Gait patterns become more pronounced with age due to weakness in muscular strength or the nervous system, making it challenging to recognize individuals who exhibit gait inconsistencies. Given the limited emphasis on enhancing gait identification for individuals exhibiting inconsistent gait patterns, this work introduces an enhanced identification technique to mitigate the impact of intra-subject gait fluctuation. The innovation aims to improve the system’s resilience by combining gait data from multiple inertial sensors and capturing spatial and temporal features. Experimental results showcase a high identification accuracy of \(97.22\%\) in the presence of pronounced gait patterns. Acknowledging the vulnerability to illicit access to gait readings and the underexplored domain of gait-based authentication security, this work introduces an authentication protocol called Gait4Auth. Gait4Auth utilizes a unique marker called a Temporary Authentication Code (TAC), which provides an additional validation layer that cross-checks a user’s identity. It adds complexity to unauthorized attempts at authentication thanks to its distributed nature. We present a lightweight liveness detection method by deploying decoy TACs that detect attempts at gaining unauthorized access.