Shape Feature Extraction

TRACE: Topological Graph Representation for Automatic Sulcal Curve Extraction

Robust Feature Extraction

[Software]

A proper geometric representation of cortical regions is a fundamental task for cortical shape analysis and landmark extraction. However, highly variable and convoluted cortical folding patterns present a significant challenge. Here, we propose a novel topological graph representation for automatic sulcal curve extraction (TRACE). During image acquisition and surface reconstruction, the reconstructed surface is susceptible to noise influences. In the presence of surface noise, TRACE identifies stable sulcal fundic regions using the line simplification method to prevent significant smoothing of the sulcal folding pattern. The Dijkstra's shortest path algorithm is then utilized to trace sulcal curves over the connected graph in the identified regions. 

Applications: Curve-based Population Analysis

Primary Sulci and Variability