Automatic Detection of Diabetes Retinopathy based on Digital Retinal Images
August 20, 2019
Classification of Thai Rice – Grain Images using Mask R – CNN and Transfer Learning
August 20, 2019

Human Gait as a Biometric for Human Re – Identification under Unconstrained Walking Conditions and Environments

Human Gait as a Biometric for Human Re – Identification under Unconstrained Walking Conditions and Environments

Faculty of Information and Communication Technology, Mahidol University

Project Title:

Gait Recognition as Human Biometric

Research Title:

Human Gait as a Biometric for Human Re-Identification under Unconstrained Walking Conditions and Environments

Researcher (s):

Asst.Prof. Dr.Worapan Kusakunniran
Mr.Lingxiang Yao
Assoc.Prof. Dr.Qiang Wu
Assoc.Prof. Dr.Jian Zhang
Prof. Dr.Zhenmin Tang
Assoc.Prof. Dr.Wankou Yang

Biometrics have been used for human verification and identification for decades. Popular human biometrics are face, fingerprint, and iris. They are effectively used in real applications under a controlled environment. Many external factors are controlled, such as lighting condition, distance to a camera, not wearing hat and sunglasses, viewing angles, and requesting physical contact. However, with these constraints, existing approaches could not be used for a purpose of automatic surveillance monitoring, where human are captured through a surveillance camera from a far distance, under any seen angles, and without any physical contact. This becomes main challenges of this research project. Thus, this project develops a new system using an alternative biometric solution using human gait.

Currently, it could achieve a perfect accuracy of 100%, but under a controlled environment and a fixed walking condition. This is not feasible to be used in a real-world scenario. This project, therefore, focuses on solving challenges of gait recognition under changes of internal and external factors such as camera viewing angles, walking directions, and walking speeds. This will make the gait-based human identification more practical to be used in real-world scenarios.

Publishing:
• L. Yao, W. Kusakunniran, Q. Wu, J. Zhang, Z. Tang, W. Yang, Robust Gait Recognition using Hybrid Descriptors based on Skeleton Gait Energy Image, Pattern Recognition Letters (PRL): Special Issue on Learning Compact Representations for Scalable Visual Recognition and Retrieval, DOI: 10.1016/j.patrec.2019.05.012
• W. Kusakunniran, Recognizing gaits on spatio-temporal feature domain, IEEE Transactions on Information Forensics and Security (TIFS), 9(9): 1416-1423, September 2014, DOI: 10.1109/TIFS.2014.2336379
• W. Kusakunniran, Attribute-based learning for gait recognition using spatio-temporal interest points, Image and Vision Computing (IVC), 32(12), 1117-1126, December 2014, DOI: 10.1016/j.imavis.2014.10.004

Award Grant related to the Project:
• Outstanding Research Award 2019, by National Research Council of Thailand (NRCT)

Key Contact Person:
Asst.Prof. Dr.Worapan Kusakunniran
Faculty of Information and Communication Technology, Mahidol University