Gait Biometric Recognition

Abstract

The gait biometric has demonstrated potential promises as an alternative or complementary identifier for use in human recognition systems. However, there is no single measure that encompasses the full set of complex dynamics reflecting what we consider to be the human gait. Instead, important aspects of gait can be measured using one or more of several analysis techniques. Among these techniques are visual approaches involving cameras, which can capture differing angles of gait from a distance, and sensor approaches, which collect information about gait while in contact with the subject being analyzed. In this chapter, we explore the ways in which these varying approaches have previously been applied to achieve gait biometric recognition, while also highlighting important possible areas of concern in their usage with respect to practicality, privacy, and security. This chapter provides a foundation of knowledge with respect to gait and machine learning, which will be built upon through the remainder of the book as we demonstrate, via the application of powerful machine learning techniques and the levels of gait recognition performance we might hope to achieve using a sensor-based approach for demonstration.