Mason J.E., Traoré I., Woungang I. Machine Learning Techniques for Gait Biometric Recognition. Using the Ground Reaction Force

Springer, 2016. — 247 p.The last two decades have seen a dramatic increase in the number of stakeholders of biometric technologies. The quality of the technologies has increased due to an improvement in underlying data processing and sensor technologies. A growing and healthy marketplace has emerged, while the number of people using, operating, or impacted by these technologies has been growing exponentially. Several new disruptive technologies have emerged, along with the diversification of the devices and platforms where biometrics are provisioned. The ubiquity of mobile phones and the multiplicity and diversity of sensors available for biometric provisioning (e.g., webcam, fingerprint reader, touchscreen, accelerometer, gyroscope, etc.) is contributing significantly to this dramatic growth of the biometric ecosystem.
Gait biometrics is one of the new technologies that have appeared in the past few decades. Gait biometric technology consists of extracting and measuring unique and distinctive patterns from human locomotion. Different forms of gait biometrics are available based on how the gait information is captured (e.g., video cameras, floor sensor, smartphones, etc.). Gait based on the ground reaction force (GRF) is the most recent form of gait biometric technology, which although lesser known than its counterparts, has shown greater promise in terms of its robustness. GRF is a measure of the force exerted by the ground back on the foot during a footstep.
The GRF-based gait biometric is the central topic of this book. Theoretical and practical underpinnings of the GRF-based gait biometric are presented in detail. The main components and processes involved in developing a GRF-based recognition system are discussed from a theoretical and experimental perspective, by revisiting existing research and introducing new results.
While the central topic of the book is GRF-based gait biometric technology, its backdrop is machine learning. Several machine learning techniques used in the literature for GRF recognition are dissected, contrasted, and investigated experimentally.
The book covers the different dimensions required for developing a GRF-based system: theoretical models, experimental models, and implementation issues. It also covers in detail several machine learning algorithms which can be used broadly for biometric recognition technologies and other similar pattern recognition problems (e.g., speech recognition). This book is intended for researchers, developers, and managers and for students of computer science and engineering, in particular graduate students at the Master’s and Ph.D. levels, working or with interest in the aforementioned areas.Introduction to Gait Biometrics
Gait Biometric Recognition
Gait Biometric Recognition Using the Footstep Ground Reaction Force
Feature Extraction
Normalization
Classification
Experimental Design and Dataset
Measured Performance
Experimental Analysis
Applications of Gait Biometrics
Conclusion and Remarks
A: Experiment Code Library