Pha Nguyen
Affiliation confirmed via AI analysis of OpenAlex, ORCID, and web sources.
Chair
Also affiliated: Ho Chi Minh City University of Science (2020); University of Arkansas System (2022–2023)
Research Areas
Links
Biography and Research Information
OverviewAI-generated summary
Pha Nguyen's research focuses on computer vision and machine learning, particularly in areas related to object tracking, activity recognition, and scene understanding in video data. Recent publications include work on self-supervised spatiotemporal transformers for group activity recognition, multi-camera multi-object tracking, and hierarchical interlacement graphs for scene graph generation. Nguyen also investigates prompt-based object tracking and hypergraph-based video scene graph generation and anticipation.
Nguyen has secured significant federal funding, serving as PI on the "Shared Arkansas Research Plan for Community Cyber Infrastructure (SHARP CCI)" from NSF for $199,592, and as Co-PI on the "E-RISE Rll: Arkansas Smart Transportation Research Incubator through Data Engineering and Science" from NSF for $7,000,000. Other federal grants include projects focused on Alaskan riverine ecosystem stability, spatial archaeometry, and lidar-based truck detection.
With an h-index of 8 and 38 total publications, Nguyen collaborates with researchers across disciplines, including Ky Luu, Page D. Dobbs, Trong-Thuan Nguyen, and Jackson Cothren, with whom they have multiple shared publications. Nguyen maintains an active lab website, indicating ongoing research activities.
Metrics
- h-index: 9
- Publications: 36
- Citations: 255
Positions
-
Amazon (United States)ORCID
-
Chair publications 2022–2025University of Arkansas at Fayetteville Listing
-
Research Engineer 2020–2021VinAI Research Department of Applied Research ORCID
Selected Publications
-
Discriminative and Generative Video Modeling (2025)Journal of the Arkansas Academy of Science OpenAlex
-
HyperGLM: HyperGraph for Video Scene Graph Generation and Anticipation (2025)
-
Autoregressive Temporal Modeling for Advanced Tracking-by-Diffusion (2025)
-
SoGAR: Self-Supervised Spatiotemporal Attention-Based Social Group Activity Recognition (2025)
-
Depth Perspective-Aware Multiple Object Tracking (2024)
-
React: recognize every action everywhere all at once (2024)
-
HAtt-Flow: Hierarchical Attention-Flow Mechanism for Group-Activity Scene Graph Generation in Videos (2024)
-
Multi-camera multi-object tracking on the move via single-stage global association approach (2024)
-
REACT: Recognize Every Action Everywhere All At Once (2024)
-
HAtt-Flow: Hierarchical Attention-Flow Mechanism for Group Activity Scene Graph Generation in Videos (2023)
-
SPARTAN: Self-supervised Spatiotemporal Transformers Approach to Group Activity Recognition (2023)
-
Self-Supervised Domain Adaptation in Crowd Counting (2022)
Federal Grants 6 $7,782,453 total
CC* CIRA: Shared Arkansas Research Plan for Community Cyber Infrastructure (SHARP CCI)
E-RISE Rll: Arkansas Smart Transportation Research Incubator through Data Engineering and Science
Collaboration Network
Top Collaborators
- SPARTAN: Self-supervised Spatiotemporal Transformers Approach to Group Activity Recognition
- Multi-camera multi-object tracking on the move via single-stage global association approach
- Self-Supervised Domain Adaptation in Crowd Counting
- React: recognize every action everywhere all at once
- SoGAR: Self-Supervised Spatiotemporal Attention-Based Social Group Activity Recognition
Showing 5 of 11 shared publications
- SPARTAN: Self-supervised Spatiotemporal Transformers Approach to Group Activity Recognition
- React: recognize every action everywhere all at once
- SoGAR: Self-Supervised Spatiotemporal Attention-Based Social Group Activity Recognition
- HAtt-Flow: Hierarchical Attention-Flow Mechanism for Group-Activity Scene Graph Generation in Videos
- REACT: Recognize Every Action Everywhere All At Once
- Multi-camera multi-object tracking on the move via single-stage global association approach
- Self-Supervised Domain Adaptation in Crowd Counting
- HAtt-Flow: Hierarchical Attention-Flow Mechanism for Group-Activity Scene Graph Generation in Videos
- HAtt-Flow: Hierarchical Attention-Flow Mechanism for Group Activity Scene Graph Generation in Videos
- SPARTAN: Self-supervised Spatiotemporal Transformers Approach to Group Activity Recognition
- React: recognize every action everywhere all at once
- SoGAR: Self-Supervised Spatiotemporal Attention-Based Social Group Activity Recognition
- HAtt-Flow: Hierarchical Attention-Flow Mechanism for Group-Activity Scene Graph Generation in Videos
- SPARTAN: Self-supervised Spatiotemporal Transformers Approach to Group Activity Recognition
- SoGAR: Self-Supervised Spatiotemporal Attention-Based Social Group Activity Recognition
- SPARTAN: Self-supervised Spatiotemporal Transformers Approach to Group Activity Recognition
- SoGAR: Self-Supervised Spatiotemporal Attention-Based Social Group Activity Recognition
- HAtt-Flow: Hierarchical Attention-Flow Mechanism for Group Activity Scene Graph Generation in Videos
- REACT: Recognize Every Action Everywhere All At Once
- Multi-camera multi-object tracking on the move via single-stage global association approach
- Depth Perspective-Aware Multiple Object Tracking
- Multi-camera multi-object tracking on the move via single-stage global association approach
- Depth Perspective-Aware Multiple Object Tracking
- Self-Supervised Domain Adaptation in Crowd Counting
- Self-Supervised Domain Adaptation in Crowd Counting
- Self-Supervised Domain Adaptation in Crowd Counting
- SPARTAN: Self-supervised Spatiotemporal Transformers Approach to Group Activity Recognition
- Multi-camera multi-object tracking on the move via single-stage global association approach
- Depth Perspective-Aware Multiple Object Tracking
Similar Researchers
Based on overlapping research topics