Welcome to Buyu Liu's Homepage

About Me

I am now a postdoc researcher in the school of informatics at the University of Edinburgh, working with Vittorio Ferrari. I was a PhD candidate in College of Engineering and Computer Science in Australian National University from 2012 to 2016, working with Xuming He (supervisor) and Stephen Gould (advisor).

I was also working in Computer Vision Group, Canberra Research Laboratory, National ICT of Australia (NICTA) .  (my previous homepage can be found here)

- Research Interests

  •   - Active Learning for Structured Output
  •   - Semantic Video/Image Parsing
  •   - Dynamic Holistic Scene Understanding
  •   - Efficient Inference and Test-time Aware Learning



G.14, School of Informatics, 10 Crichton St, Edinburgh EH8 9AB


Buyu dot Liu at anu dot edu dot au   


Buyu dot Liu at ed dot ac dot uk (Preferred)


  • We have one paper accepted in ICCV17 --- Active Learning for Human Pose Estimation
  • We have one paper accepted in ECCV16 --- Learning Dynamic Hierarchical Models for Anytime Scene Labeling
  • We have uploaded results on test videos of CamVid. More details can be found in publication page.
  • We have one paper accepted in CVPR15 --- Multiclass Semantic Video Segmentation with Object-level Active Inference.
  • We have one paper accepted in WACV15. 
  • Location

    Locations of visitors to this page
Subpages (2): Publication Research