A5021022119- Gait Recognition and Analysis
- Indoor and Outdoor Localization Technologies
- Human Pose and Action Recognition
- Diabetic Foot Ulcer Assessment and Management
- Fluid Dynamics and Turbulent Flows
- Microgrid Control and Optimization
- stochastic dynamics and bifurcation
- Meteorological Phenomena and Simulations
- Neural Networks Stability and Synchronization
- Advanced Neural Network Applications
- Scientific Research and Discoveries
- Video Surveillance and Tracking Methods
- Neural Networks and Applications
- Multilevel Inverters and Converters
- Theoretical and Computational Physics
- Wind and Air Flow Studies
- Advanced DC-DC Converters
Yunnan University
2024
China University of Mining and Technology
2024
Shanghai Jiao Tong University
2024
China Southern Power Grid (China)
2022
State Key Laboratory of Virtual Reality Technology and Systems
2021
Beihang University
2021
In this work we propose to boost video-based person re-identification (Re-ID) by using attribute-enhanced feature presentation. To end, not only try use the ID-relevant attributes more effectively, but also for first time in literature harness ID-irrelevant help model training. The former mainly include gender, age, clothing characteristics, etc., which contain rich and supplementary information about pedestrian; latter viewpoint, action, are seldom used identification previously....
Gait recognition under multiple views is an important computer vision and pattern task. In the emerging convolutional neural network based approaches, information of view angle ignored to some extent. Instead direct estimation training view-specific models, we propose a compatible framework that can embed into existing architectures gait recognition. The embedding simply achieved by selective projection layer. Experimental results on two large public datasets show proposed very effective.
Gait is an important biometric that can recognize people at a distance. Recently, Disentangled Representation Learning (DRL) has been introduced for distinguishing identity-irrelevant covariate features from identity better recognition performance. However, such simple gait energy image (GEI) pairing operation inevitably brings in over-disentanglement effects degrade the To address this issue, we proposed feature control gate module compensates discriminative loss by using additional...
For continuum fields such as turbulence, analyses of the field structure offer insights into their kinematic and dynamic properties. To ensure are quantitative rather than merely illustrative, two conditions essential: space-filling quantification. A pertinent example is dissipation element (DE) structure, which however susceptible to noisy interference, rendering it inefficient for extracting large-scale features field. In this study, multi-level DE proposed based on extremal point concept....
Field variables of a complex system will essentially inherit the kinematic and dynamic properties system. This paper focuses on identification scalar field structure based extremal points, from which statistical features can be quantitatively interpreted. In more general case, points defined at multiple scale levels, depending their validity ranges. Accordingly, multi-level dissipation element (DE) analysis has been developed by decomposing entire into space-filling units different levels....
Gait recognition under multiple views is an important computer vision and pattern task. In the emerging convolutional neural network based approaches, information of view angle ignored to some extent. Instead direct estimation training view-specific models, we propose a compatible framework that can embed into existing architectures gait recognition. The embedding simply achieved by selective projection layer. Experimental results on two large public datasets show proposed very effective.