Graphics Visualisation and Modelling of 3D Datasets

Overview

Scientific simulations and instruments like MRI or CT-scans produce nowadays a large amount of data that can be hard for scientists to interpret. The use of advanced Graphics and Visualisation algorithms possibly supported by AI can allow people to get a better understanding of the data. This group is particularly concerned with efficient processing of 3D datasets that can be encoded for instance as a 3D mesh or rectilinearly sampled grids.

This group has successfully looked at various problems in recent years, including issues related to dealing with large amount of data while still providing real-time processing and/or display of images to users. Members of this group have for instance worked on new algorithms having a better complexity to render 3D datasets or proposed better data encoding data to meet these goals. The use of AI to provide state-of-the-art lossless compression of large medical datasets was for instance pioneered within the group.

This group is also interested in various aspects of modelling Physical phenomena used in 3D renderings, including real-time visualisation of incompressible fluids or using bonded discrete element methods for simulating fractures.

Simulation of incompressible fluids

Lossless compression of medical data using an LSTM network to predict the next voxel value

3D display of the classical “Stanford bunny”

Our Facilities

Key Enabling Technologies:

Specialist Area Leads

Events