Left Atrial Segmentation Challenge: A Unified Benchmarking Framework [chapter]

Catalina Tobon-Gomez, Jochen Peters, Juergen Weese, Karen Pinto, Rashed Karim, Tobias Schaeffter, Reza Razavi, Kawal S. Rhode
2014 Lecture Notes in Computer Science  
The knowledge of left atrial (LA) anatomy is important for atrial fibrillation ablation guidance. More recently, LA anatomical models have been used for cardiac biophysical modelling. Segmentation of the LA from Magnetic Resonance Imaging (MRI) and Computed Tomography (CT) images is a complex problem. We aimed at evaluating current algorithms that address this problem by creating a unified benchmarking framework through the mechanism of a challenge, the Left Atrial Segmentation Challenge 2013
more » ... ASC'13). Thirty MRI and thirty CT datasets were provided to participants for segmentation. Ten data sets for each modality were provided with expert manual segmentations for algorithm training. The other 20 data sets per modality were used for evaluation. The datasets were provided by King's College London and Philips Technologie GmbH. Each participant segmented the LA including a short part of the LA appendage trunk plus the proximal parts of the pulmonary veins. Details on the evaluation framework and the results obtained in this challenge are presented in this manuscript. The results showed that methodologies combining statistical models with region growing approaches were the most appropriate to handle the proposed task.
doi:10.1007/978-3-642-54268-8_1 fatcat:ejttwrw4yvd2jknagtlydalrre