GaitSet: Regarding Gait as a Set for Cross-View Gait Recognition

Chao, Hanqing; He, Yejun; Zhang, Junping; Feng, Jianfeng

doi:10.1609/aaai.v33i01.33018126

Cited by 462 publications

(501 citation statements)

References 7 publications

Supporting

Mentioning

449

Contrasting

Unclassified

Order By: Relevance

“…Zhang et al [38] used the KPCA-LPP method and achieved 91.12% average gait recognition accuracy. Chao et al [39] achieved 95% average gait recognition accuracy using the method of regarding the gait as a set. Wolf et al [40] used the 3DCNN method with average gait recognition accuracy of 97.35%.…”

Section: Methodsmentioning

confidence: 99%

Gait Recognition Based on the Feature Extraction of Gabor Filter and Linear Discriminant Analysis and Improved Local Coupled Extreme Learning Machine

Guo

Zhang

et al. 2020

Mathematical Problems in Engineering

View full text Add to dashboard Cite

A gait energy image contains much gait information, which is one of the most effective means to recognize gait characteristics. The accuracy of gait recognition is greatly affected by covariates, such as the viewing angle, occlusion of clothing, and walking speed. Gait features differ somewhat by angles. Therefore, how to improve the recognition accuracy of a cross-view gait is a challenging task. This study proposes a new gait recognition algorithm structure. A Gabor filter is used to extract gait features from gait energy images, since it can extract features of different directions and scales. We use linear discriminant analysis (LDA) to tackle the problem that the feature dimension restricts the process. Finally, the improved local coupled extreme learning machine based on particle swarm optimization is used for the classification process of the extracted features of the gait. The proposed method and other current mainstream algorithms are compared in terms of the recognition accuracy based on the CASIA-A and CASIA-B datasets, and the simulation results show that the proposed algorithm has good performance and performs well at cross-view gait recognition.

show abstract

Section: Methodsmentioning

confidence: 99%

Gait Recognition Based on the Feature Extraction of Gabor Filter and Linear Discriminant Analysis and Improved Local Coupled Extreme Learning Machine

Guo

Zhang

et al. 2020

Mathematical Problems in Engineering

View full text Add to dashboard Cite

show abstract

“…Ben et al [29], [30] proposed another two cross-view gait recognition methods respectively based on matrix and tensor, which are suitable for reducing the small sample size problem in discriminative subspace selection. Some other cross-view gait recognition methods can be found in [31], [32], [33]. To extract invariant gait feature, generative adversarial networks (GAN) are also employed in [34], [35].…”

Section: Gait Recognitionmentioning

confidence: 99%

Performance Evaluation of Model-Based Gait on Multi-View Very Large Population Database With Pose Sequences

Makihara

et al. 2020

IEEE Trans. Biom. Behav. Identity Sci.

121

View full text Add to dashboard Cite

Model-based gait recognition is considered to be promising due to the robustness against some variations, such as clothing and baggage carried. Although model-based gait recognition has not been fully explored due to the difficulty of human body model fitting and the lack of a large-scale gait database, recent progress in deep learning-based approaches to human body model fitting and human pose estimation is mitigating the difficulty. In this paper, we, therefore, address the remaining issue by presenting a large-scale human pose-based gait database, OUMVLP-Pose, which is based on a publicly available multi-view large-scale gait database, OUMVLP. OUMVLP-Pose has many unique advantages compared with other public databases. First, OUMVLP-Pose is the first gait database that provides two datasets of human pose sequences extracted by two standard deep learning-based pose estimation algorithms, OpenPose and AlphaPose. Second, it contains multi-view large-scale data, i.e., over 10,000 subjects and 14 views for each subject. In addition, we also provide benchmarks in which different kinds of gait recognition methods, including model-based methods and appearance-based methods, have been evaluated comprehensively. The model-based gait recognition methods have shown promising performances. We believe this database, OUMVLP-Pose, will greatly promote model-based gait recognition in the next few years.

show abstract

“…To further evaluate the performance of GRU-GSFSL under multi-views with various factors, several related methods are compared using CASIA Dataset B, i.e., VI-MGR [50], GFI-CCA [48], RLTDA [51], GEI-SVD [47], Robust VTM [52], GEI-GaitSet [4], FT-SVD [53], and HBPS-GLM [54]. The performances of the related methods are extracted from their paper under different views with various factors, i.e., bag carrying or clothes changes.…”

Section: Multi-view Gait Recognition Under Various Conditionsmentioning

confidence: 99%

Gait Recognition and Understanding Based on Hierarchical Temporal Memory Using 3D Gait Semantic Folding

Luo

Tjahjadi

2020

Sensors

View full text Add to dashboard Cite

Gait recognition and understanding systems have shown a wide-ranging application prospect. However, their use of unstructured data from image and video has affected their performance, e.g., they are easily influenced by multi-views, occlusion, clothes, and object carrying conditions. This paper addresses these problems using a realistic 3-dimensional (3D) human structural data and sequential pattern learning framework with top-down attention modulating mechanism based on Hierarchical Temporal Memory (HTM). First, an accurate 2-dimensional (2D) to 3D human body pose and shape semantic parameters estimation method is proposed, which exploits the advantages of an instance-level body parsing model and a virtual dressing method. Second, by using gait semantic folding, the estimated body parameters are encoded using a sparse 2D matrix to construct the structural gait semantic image. In order to achieve time-based gait recognition, an HTM Network is constructed to obtain the sequence-level gait sparse distribution representations (SL-GSDRs). A top-down attention mechanism is introduced to deal with various conditions including multi-views by refining the SL-GSDRs, according to prior knowledge. The proposed gait learning model not only aids gait recognition tasks to overcome the difficulties in real application scenarios but also provides the structured gait semantic images for visual cognition. Experimental analyses on CMU MoBo, CASIA B, TUM-IITKGP, and KY4D datasets show a significant performance gain in terms of accuracy and robustness.

show abstract

GaitSet: Regarding Gait as a Set for Cross-View Gait Recognition

Cited by 462 publications

References 7 publications

Gait Recognition Based on the Feature Extraction of Gabor Filter and Linear Discriminant Analysis and Improved Local Coupled Extreme Learning Machine

Gait Recognition Based on the Feature Extraction of Gabor Filter and Linear Discriminant Analysis and Improved Local Coupled Extreme Learning Machine

Performance Evaluation of Model-Based Gait on Multi-View Very Large Population Database With Pose Sequences

Gait Recognition and Understanding Based on Hierarchical Temporal Memory Using 3D Gait Semantic Folding

Contact Info

Product

Resources

About