Research on fastDTW and isolation forest-based lightweight gesture authentication

Abstract

This study explores the use of the accelerometer and gyroscope sensors in an Android mobile phone to capture data during user gesture performance. The recorded data is then processed and analyzed to extract feature values. The authentication of the user's identity may be achieved by this method, which is characterized by minimal constraints. This study initially compares the user's gesture to the template gesture using FastDTW. After that, an adaptive weighted vote determines the closest gesture template. This checks if the user is doing the template category correctly. The second stage extracts effective feature values from authentic user gestures. These feature values include efficient time-frequency domain feature values and the data extreme point spacing-to-length ratio. After that, authentic gestures are used to build the Isolation Forest to verify the user. This approach suggests a reduction in FRR, FAR, and an increase in accuracy, while using a smaller amount of data.

KEYWORDS:

• Gesture recognition
• identity authentication
• mobile phone safety
• FastDTW
• isolation Forest

Acknowledgment

This work is supported by the Radio, television and network audiovisual medium and long-term science and technology plan: soft terminal key technology and ecological research(No. ZG23001).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

HanFan Liu

HanFan Liu received the BEng degree from Communication University of China in 2023.He is currently a master student at School of Information and Communication Engineering, Communication University of China. His main research interest is Intelligent Networks and Big Data.

Pei Tian

Pei Tian received his BEng degree from Northwestern Polytechnical University in 1991, MEng degree from Northwestern Polytechnical University in 1998, and DEng degree from Beihang University in 2004. He is currently a professor and head of the Department of Electronic Information Engineering, School of Information Engineering, Communication University of China.

His main research areas are software development and testing technology, artificial intelligence technology, etc. He has hosted and participated in more than ten national and provincial projects, and has published more than twenty EI/SCI retrieved papers and four monographs in recent years. He has been awarded the second prize of National Defense Science and Technology and the Science and Technology Innovation Award of the General Administration of Radio and Television.

HaiFang Yao

HaiFang Yao received the BEng degree from Communication University of China in 2021.She is currently a master student at School of Information and Communication Engineering, Communication University of China. Her main research interest is Intelligent Networks and Big Data.

Cong Shen

Cong Shen received the BEng degree from Chengdu University of Technology in 2022.He is currently a master student at School of Information and Communication Engineering, Communication University of China. His main research interest is Intelligent Networks and Big Data.

FengYang Hu

FengYang Hu received the BEng degree from Zhejiang University of Technology in 2022.He is currently a master student at School of Information and Communication Engineering, Communication University of China. His main research interest is Intelligent Networks and Big Data.