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.

Clothing invariant human gait recognition using modified local optimal oriented pattern binary descriptor

Clothing-invariant human gait recognition using an adaptive outlier detection method

Covariate conscious approach for Gait recognition based upon Zernike moment invariants

Human Gait Recognition Based on Self-Adaptive Hidden Markov Model

Human Gait Recognition by Integrating Motion Feature and Shape Feature

Gait Recognition Based on Local Graphical Skeleton Descriptor with Pairwise Similarity Network.

Human Identification Based on Gait Manifold

Non-local Gait Feature Extraction and Human Identification

Human Gait Recognition using Machine Learning Teachnolgies for Inclusive Innovation

Part-based Human Gait Identification under Clothing and Carrying Condition Variations

GaitLRDF: gait recognition via local relevant feature representation and discriminative feature learning

Persistent-homology-based gait recognition

Enhancing gait recognition by multimodal fusion of mobilenetv1 and xception features via PCA for OaA-SVM classification

Feature integration for frontal gait recognition through contour image analysis

Gabor-Based Discriminative Common Vectors for Gait Recognition

Gait recognition via independent component analysis based on support vector machine and neural network

Gait Recognition for Human Identification Based on Ica and Fuzzy Svm Through Multiple Views Fusion

Sparse Tensor Discriminative Locality Alignment For Gait Recognition

Intra-class Variations with Deep Learning-Based Gait Analysis: A Comprehensive Survey of Covariates and Methods

Collaborative Feature Learning for Gait Recognition under Cloth Changes

Condition-Aware Comparison Scheme for Gait Recognition