Human Identification Using Gait Skeletal Joint Distance Features

Human Identification Using Gait Skeletal Joint Distance Features

Md Wasiur Rahman (Department of Computer Science, University of Calgary, Calgary, Canada) and Marina L. Gavrilova (Department of Computer Science, University of Calgary, Calgary, Canada)
DOI: 10.4018/IJSSCI.2017100102
OnDemand PDF Download:
No Current Special Offers


Gait not only defines the way a person walks, but also provides insights on an individual's daily routine, mental state or even cognitive function. The importance of incorporating cognitive behavior and analysis in biometric systems has been noted recently. In this article, authors develop a biometric security system using gait-based skeletal information obtained from Microsoft Kinect v1 sensor. The gait cycle is calculated by detecting the three consecutive local minima between the joint distance of left and right ankles. Authors have utilized the distance feature vector for each of the joints with respect to other joints in the gait cycle. After mean and variance features are extracted from the distance feature vector, the KNN algorithm is used for classification purpose. The classification accuracy of the authors' approach is 93.33%. Experimental results show that the proposed approach achieves better recognition accuracy then other state-of-the-art approaches. Incorporating gait biometric in a situation awareness system for identification of a mental state is one of the future directions of this research.
Article Preview


Biometric technologies are considered to be a more convenient and secure way of user authentication, as compared to traditional identification and verification methods (ID, cards or passwords). It is well known that human biometric identifiers can be divided into two categories: physiological biometrics, which may relate to parts of the body (face, ear, palm, fingerprint, iris), and behavioral biometrics including voice, signature, handwriting and gait (Gavrilova, & Monowar, 2012). There is also an emerging area of biometric domain, that uses social on-line interactions and aesthetics for biometric recognition (Sultana, Paul, & Gavrilova, 2017; Azam, & Gavrilova, 2017). Some biometric identifiers require person cooperation because it is difficult to obtain a person’s face, fingerprint, iris or voice from a location (Munsell, Tumlyakov, & Wang, 2017). Some others biometric identifiers require high quality of an image or a video for accurate person identification and feature extraction. Gait is one of the few biometric identifiers where person cooperation is not needed, and person can be identified using even a low-quality image or a video (Das, Guang, & Cheng-Tsun, 2014). Thus, the recent popularity of gait biometric can be attributed to its unobtrusiveness, universality and non-vulnerability in the case of a spoof attack, as gait is difficult to hide or imitate. As a result, gait analysis has been conducted in various applications such as video surveillance systems (Cucchiara,Grana, Pretti, & Vezzani, 2005), 3D human body modeling (Bae, & Park, 2013), forensic science (Bouchrika, Goffredo, Carter, & Nixon, 2011), and elderly population health assessment (Kressing, & Beauchet, 2006).

Gait can not only define the way a person walks, but also provides interesting cues on individuals daily routine, mental state, health state or even cognitive function (Wang, 2011). The importance of incorporating cognitive behavior and analysis in biometric systems has been noted recently (Wang, Widrow, Zadeh, Howard, Wood, Chan, & Gavrilova, 2016). One of the early works conducted in BT lab demonstrated that combining auxiliary cues from the gait videos in addition to extracting traditional gait features can significantly enhance the accuracy of subject identification (Bazazian, & Gavrilova, 2015). An interesting link between gait and cognition has been observed recently (Wang, Fariello, Gavrilova, Kinsner, & Shell, 2013). Normally, gait and cognitive function of a person are being evaluated independently; however, the strong correlation between changes in walking style, for instance, and some mental illnesses or cognitive impairments have been noted and investigated (Montero-Obasso, Verghese, Beauchet, & Hausdroff, 2012). Gait skeletal information can not only be used for person identification, but also for other applications, including gender detection (Kastaniotis, Theodorakopoulos, Theoharatos, Economou, & Fotopoulos, 2015). It can also be used to recognize = actions of a person from recorded sequences or in real-time (Ahmed, Paul, & Gavrilova, 2016).

Complete Article List

Search this Journal:
Open Access Articles
Volume 14: 4 Issues (2022): 1 Released, 3 Forthcoming
Volume 13: 4 Issues (2021)
Volume 12: 4 Issues (2020)
Volume 11: 4 Issues (2019)
Volume 10: 4 Issues (2018)
Volume 9: 4 Issues (2017)
Volume 8: 4 Issues (2016)
Volume 7: 4 Issues (2015)
Volume 6: 4 Issues (2014)
Volume 5: 4 Issues (2013)
Volume 4: 4 Issues (2012)
Volume 3: 4 Issues (2011)
Volume 2: 4 Issues (2010)
Volume 1: 4 Issues (2009)
View Complete Journal Contents Listing