Abstract:Gait-based pedestrian identification has important applications in intelligent surveillance. From anatomical viewpoint, the physical uniqueness of human gait is physiological discriminative of individuals. Therefore, in theory, like fingerprint and face, gait is used as a biometric for pedestrian identification. However, gait-based pedestrian identification still faces multiple challenges due to a vast diversity of walking conditions and complex acquisition environment during data collection. In this paper, through combining the nonlinear dimensionality reduction by using gait manifold and the temporal feature of gated recurrent unit (GRU) together, we propose a novel gait-based pedestrian identification framework. Firstly, we design a temporal enhancement module to construct a series of frame-by-frame gait trend energy images (ff-GTEIs), which represents spatiotemporal gait characteristics and does not reduce the number of samples. Secondly, a supervised locally linear embedding (LLE) dimensionality reduction scheme is proposed, which generates a low-dimensional gait manifold for each pedestrian and transforms all ff-GTEIs into corresponding gait manifold space. Thirdly, a new pedestrian identification network based on residual GRU is proposed, which is able to identify a person by comprehensively considering the similarity between its gait and corresponding gait manifold. Finally, a series of comparative experiments are carried out based on well-known gait datasets, the experimental results show that the proposed framework for pedestrian recognition in this paper exceeds most existing methods, and has achieved an average correct recognition rate 97.4% and 99.6% based on the open-accessed gait dataset CASIA Dataset B and OU-ISIR LP dataset, respectively.

Recognition of human gaits

Gait analysis of human locomotion based on motion capture system

Accelerometer-based Gait Recognition Via Voting by Signature Points

A Gait Recognition System for Interaction with a Homecare Mobile Robot

Hybrid Silhouette-skeleton Body Representation for Gait Recognition

Gait-Event-Based human intention recognition approach for lower limb

Human Gait Recognition Based on Self-Adaptive Hidden Markov Model

Classification and recognition of dynamical models: the role of phase, independent components, kernels and optimal transport

Human Identification Based on Gait Manifold

Gaitor: Learning a Unified Representation Across Gaits for Real-World Quadruped Locomotion

Human Gait Recognition by Integrating Motion Feature and Shape Feature

Human Gait Recognition Based on Frame-by-Frame Gait Energy Images and Convolutional Long Short-Term Memory

Human Identity and Gender Recognition From Gait Sequences With Arbitrary Walking Directions

Pose-based Deep Gait Recognition

End-to-End Model-Based Gait Recognition.

Human Gait Recognition using Machine Learning Teachnolgies for Inclusive Innovation

Human Gait Recognition with Matrix Representation

Static and Dynamic Features Analysis from Human Skeletons for Gait Recognition

Persistent-homology-based gait recognition

Automatic gait recognition from a distance

[A Study of Gait Recognition Based on Kinematics and Kinetics Parameters].