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.

Gait Authentication Based on Acceleration Characteristic Point Extraction

Accelerometer-based Gait Recognition Via Voting by Signature Points

Estimating Walking Distance Based on Single Accelerometer

Video2mmPoint: Synthesizing Mmwave Point Cloud Data from Videos for Gait Recognition

Accelerometer-Based Gait Recognition by Sparse Representation of Signature Points with Clusters

Formalizing PQRST Complex in Accelerometer-based Gait Cycle for Authentication

Human Identification Based on Gait Manifold

Non-local Gait Feature Extraction and Human Identification

Extracting Human Gait Series Based on the Wavelet Transform

Gait Pattern Recognition Using Accelerometers

A gait stability evaluation method based on wearable acceleration sensors

Algorithm for Gait Parameters Estimation Based on Heel-Mounted Inertial Sensors

An Iterative Gait Prototype Learning Algorithm Based on Tangent Distance

Accelerometer-Based Gait Segmentation: Simultaneously User and Adversary Identification

Inertial Measurement Unit-Based Real-Time Adaptive Algorithm for Human Walking Pattern and Gait Event Detection

A Gait Cadence Measurement Method Based on Match Filter and Turns Detection for Human Locomotion Monitoring.

DyGait: Exploiting Dynamic Representations for High-performance Gait Recognition

Advanced Homological Analysis for Biometric Identification Using Accelerometer

Learning Temporal Attention Based Keypoint-Guided Embedding for Gait Recognition

Wearable Device-Based Gait Recognition Using Angle Embedded Gait Dynamic Images and a Convolutional Neural Network

Gait Classification and Identity Authentication Using CNN