Restoration and Conformal Mapping of Triangulated Surfaces

The proposed algorithm is thought as a pre-processing tool for inflating cortical surface meshes, which have been created using segmentation and subsequent triangulation of magnetic resonance images (MRI). It works directly on the triangulated surface and therefore is completely independent from the underlying segmentation. It needs no further information than the triangle mesh itself, which makes it generally applicable for the removal of topological noise. The presented approach guarantees a manifold mesh by locally examining connectivity in the neighbourhood of each vertex and removing non-manifold components.

Comparision between initial ((a), (b)) and inflated ((b), (c)) representation of the cortical surface. The colouring encodes the curvature; sulci are coloured red and gyri are coloured green

Former Project Members


Compumedics Neuroscan is a developer of software and hardware for EEG, ERPs, Source Localization and Functional Neuroimaging. The presented algorithms were implemented during an internship at this very company.