Thanh-Dat Truong

Postdoctoral Fellow

University of Arkansas at Fayetteville

postdoc

11 h-index 60 pubs 397 cited

Research Areas

Advanced Neural Network Applications Domain Adaptation Action Recognition Semantic Scene Segmentation Movement Analysis Face Recognition Computer Vision

Links

ORCID

Is this your profile? Verify and claim your profile

OverviewAI-generated summary

Thanh-Dat Truong's research centers on the application of advanced machine learning techniques, particularly neural networks, for complex pattern recognition and domain adaptation tasks. His work has explored robust action recognition using directed attention in Transformers, and semantic scene segmentation through bijective maximum likelihood approaches and fairness-oriented domain adaptation. He has also investigated graph convolutional neural networks for movement analysis in neurological and musculoskeletal disorders, and developed lightweight attentive distillation for age-invariant face recognition. Truong's publications include work on self-supervised domain adaptation in crowd counting and optimal transport-based methods for unsupervised domain adaptation. His research also extends to foundation models for visual understanding, such as a large-scale dataset for insect recognition.

Metrics

  • h-index: 11
  • Publications: 60
  • Citations: 397

Selected Publications

  • DEGA: Dynamic Entropy Guided Adaptation (2025) DOI
  • Insect-Foundation: A Foundation Model and Large Multimodal Dataset for Vision-Language Insect Understanding (2025) DOI
  • Cross-view action recognition understanding from exocentric to egocentric perspective (2024) DOI
  • CONDA: Continual Unsupervised Domain Adaptation Learning in Visual Perception for Self-Driving Cars (2024) DOI
  • Insect-Foundation: A Foundation Model and Large-Scale 1M Dataset for Visual Insect Understanding (2024) DOI
  • LIAAD: Lightweight attentive angular distillation for large-scale age-invariant face recognition (2023) DOI
  • Self-Supervised Domain Adaptation in Crowd Counting (2022) DOI
  • EQAdap: Equipollent Domain Adaptation Approach to Image Deblurring (2022) DOI
  • BiMaL: Bijective Maximum Likelihood Approach to Domain Adaptation in Semantic Scene Segmentation (2021) DOI
  • Movement Analysis for Neurological and Musculoskeletal Disorders Using Graph Convolutional Neural Network (2021) DOI
  • Fast Flow Reconstruction via Robust Invertible n × n Convolution (2021) DOI
  • DyGLIP: A Dynamic Graph Model with Link Prediction for Accurate Multi-Camera Multiple Object Tracking (2021) DOI