Segmentation of Airways Based on Gradient Vector Flow
|Authors||Bauer Christian, Bischof Horst, Beichel Reinhard|
The Second International Workshop on Pulmonary Image Analysis, Proc. of Medical Image Computing and Computer Assisted Intervention
We present an automated approach for the segmentation of airways in CT datasets. The approach utilizes the Gradient Vector Flow and consists of two main processing steps. Initially, airway-like structures are identified and their centerlines are extracted. These centerlines are used in a second step to initialize the actual segmentation of the corresponding airways. An evaluation on 20 clinical datasets shows that our method achieves a good average airway branch count (63.0\%) without any major leakage.