Papers

19

Total Citations

827

H-Index

10

About

Siyu Tang is a prominent computer vision and machine learning researcher whose work spans gaze estimation, 3D human body understanding, human-scene interaction, and robotic grasping. Based at ETH Zurich, Tang has made significant strides in developing datasets and methods that push the boundaries of how machines perceive and model human behavior in complex, real-world environments. Among her most influential contributions is the ETH-XGaze dataset (2020, 297 citations), a large-scale benchmark for gaze estimation under extreme head pose and gaze variation that has become a key reference for the research community. Her work on "Grasping Field" (2020, 190 citations) introduced novel implicit representations for synthesizing realistic human grasps, addressing the longstanding challenge of modeling the hand's many degrees of freedom. Further extending her reach into embodied AI, Tang has tackled 4D human body capture in 3D scenes, egocentric body estimation via head-mounted devices, and stochastic whole-body grasping through the SAGA framework. Her contributions to 3D point cloud segmentation and fine-grained action parsing round out a research portfolio that bridges perception, synthesis, and interaction — making her work essential reading for anyone working at the intersection of computer vision, robotics, and human-centered AI.

Research Focus

Key Achievements

10
H-Index
19
Papers
827
Total Citations
44
Avg Citations/Paper
🏆 Most Cited Paper
ETH-XGaze: A Large Scale Dataset for Gaze Estimation Under Extreme Head Pose and Gaze Variation
297 citations · 2020
📈 Most Prolific Year: 2020 (4 Papers)
🤝 Key Collaborators: 40
🏛 Institutions: ETH Zurich, University of Tübingen, Shanghai Ship and Shipping Research Institute, Max Planck Society

Top Papers

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 15 days ago