Example PhD

Regularity-based Editing of Point Models

Supervisor: Dr F.C. Langbein

Keywords: Point-based modelling, discrete symmetries, pattern recognition, geometric modelling

The introduction of 3D scanners has enabled simple, rapid creation of 3D models from real-world objects, which often consist of a large amount of data giving the positions and colours of sampled points. Instead of creating polygonal meshes, boundary representations or other higher-level models from the sampled points, an alternative approach is to describe the shape of the object with the sampled points directly. Compared to the other representations the simplicity and flexibility of such point models makes them an attractive choice for efficient geometry processing and rendering algorithms, in particular on parallel hardware. Although there have been significant advances in rendering and processing such point models, editing them in a meaningful way beyond 3D point operations remains difficult. The difficulties here are similar to situations where a user wishes to edit the contents a 2D digital image using pixel-based operations. Converting the point model into a different representation that makes higher-level structures of the shape apparent also means the flexibility of the point representation is lost and one has to work with a single higher-level interpretation of the shape. To preserve the flexibility of the point models this project aims to devise efficient shape analysis algorithms that operate directly on the point cloud to quickly detect regular geometric patterns which in turn can be used immediately to edit the model.

Key Skills/Background: requires strong mathematical and computing skills

Contact: Dr F.C. Langbein to discuss this research topic.