Skip to main content

Visual computing

We work at the intersection of computer vision and computer graphics.

Who we are

Find out more about us and join us at one of our seminars.

Join us

Find out how to join us as a PhD student.

Research links

Our group focuses on human-centric visual computing, which comprises work across computer vision and computer graphics with cross-over into artificial intelligence, machine learning, education, healthcare, human-computer interaction and entertainment applications.

We have links across the University, including:

Motion capture studio

Research topics

  • Learning models of shape, appearance and dynamics from images and video.
  • Facial analysis and synthesis, performance capture and animation, applied perception for vision and graphics
  • Large scale machine learning and data science; theoretical and applied to medical imaging and computer vision
  • Machine learning applied to projects from handwriting, multi-modal camera pose estimation to finding pulsars from LIGO data
  • Automatically processing photos and video into art, and teaching computers to recognise things as depicted in art
  • Development of general autonomous systems and their applications in manufacturing and professional capture
  • Shape analysis and optimization, image/shape/scene synthesis, computational design and fabrication, interactive techniques based on VR/AR
  • Unsupervised learning and data-driven graphic design with minimal annotated data
  • Visual recognition with scarce supervision, multi-modal visual representations and generative modelling
a dog wearing a motion capture suit

Research projects

  • Generative models for shape and appearance
  • Data efficient machine learning
  • Uncertainty in computer vision
  • Intelligent software tools to assist users from other domains (e.g. artists or clinicians)
  • Deep reinforcement learning for model predictive control
  • Autonomous filming and manufacturing
  • Film style synthesis
  • Dynamic visual SLAMs
  • Indoor scene understanding and synthesis
  • Shape abstraction based content generation
  • Structure/geometry-aware image synthesis
  • Developable surface and paper craft modelling
  • Virtual scene synthesis and navigation
  • Audio-visual representation learning for assistive applications
  • Generative modeling as a means for evaluation
  • Classifying images regardless of depictive style
  • Acquiring complete and editable outdoor models from video and images
A woman blindfolded with earphones in holding a screen

Research outputs

Take a look at recent papers, articles and conference contributions from our staff and students on Bath research portal.