Robust Gait-Based Person Identification against Walking Speed Variations

Abstract

SUMMARYVariations in walking speed have a strong impact on gaitbased person identification. We propose a method that is robust against walking-speed variations. It is based on a combination of cubic higherorder local auto-correlation (CHLAC), gait silhouette-based principal component analysis (GSP), and a statistical framework using hidden Markov models (HMMs). The CHLAC features capture the within-phase spatiotemporal characteristics of each individual, the GSP features retain more shape/phase information for better gait sequence alignment, and the HMMs classify the ID of each gait even when walking speed changes nonlinearly. We compared the performance of our method with other conventional methods using five different databases, SOTON, USF-NIST, CMU-MoBo, TokyoTech A and TokyoTech B. The proposed method was equal to or better than the others when the speed did not change greatly, and it was significantly better when the speed varied across and within a gait sequence.

Gait Recognition under Speed Transition

“…In spite of considerable progress in gait recognition research, invariance to large intra-class variations of gait features is still a challenging problem (e.g., walking speed [33,35,2], viewpoint [23,15,27], clothing [12], and elapsed time [31,29]). Among these features, speed invariance is one of the most important issues because walking speed often varies depending on situation.…”

“…Walking speed change is considered from two aspects: (1) walking speed change between a matching pair of gait image sequences and (2) walking speed change within a gait image sequence. While most studies of speed-invariant gait recognition [33,35,2] consider the first aspect as cross-speed gait recognition, so far very limited attention has been paid to the second aspect.…”

“…If the speed transition is only derived from the first factor, the problem can be solved by non-linear time warping techniques [6,37,2]. However, both factors simultaneously change under a real speed transition, and hence rate adjustment alone does not work for the speed-transited gait image sequence.…”

Statistical Person Verification Using Behavioral Patterns from Complex Human Motion

“…However, they cannot be changed if the biometric data is counterfeited. In behavioral biometrics, identity is verified through action patterns which can be repeated in a unique manner, such as voice [5] and gait [1]. Behavioral biometrics are less stable since behavior may change due to the physical state of the individual.…”

A statistical approach for person verification using human behavioral patterns

“…However, templates are weak against the natural variations of individual behavioral patterns since they only encode an average representation of observed samples [19]. Statistical models such as Gaussian mixture models (GMM) and hidden Markov models (HMM) have successfully handled variations in individual behavioral patterns [20][21][22][23]. For example, Kale et al [21] showed that HMMs were more robust than templates for gait recognition.…”