Van-Hung Le
Papers
6
Total Citations
76
H-Index
3
About
Van-Hung Le is a computer vision and deep learning researcher whose work centers on human pose estimation, hand gesture recognition, and robot perception systems. His most influential contributions lie in developing efficient, real-time frameworks for understanding human body and hand movements, with significant implications for sports analytics, healthcare, robotics, and human-computer interaction. Le's landmark 2022 paper introducing the YOLOv5-HR-TCM framework for unified 2D/3D human pose estimation has garnered 33 citations, demonstrating its impact by addressing the critical gap between accuracy and real-time performance in pose estimation pipelines. A companion study combining YOLOv5 and HRNet for 2D keypoint detection accumulated 20 citations, further establishing his expertise in high-accuracy pose analysis. His 2020 survey on 3D hand skeleton estimation via convolutional neural networks, cited 17 times, reflects his broader commitment to synthesizing knowledge across emerging subfields. More recently, Le has contributed a hand gesture recognition dataset (TQU-HG) and explored Visual SLAM systems using RGB-D imagery, demonstrating an expanding research scope toward embodied AI and assistive robotics. Across his body of work, Le consistently bridges theoretical rigor with practical deployment, making him a valuable voice in applied computer vision research.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 53D Object Finding Using Geometrical Constraints on Depth Images2 citations · 2015
- 6