Evaluation of Methods for Pulmonary Image Registration 2010
Registration of thoracic CT data is one of the most common areas of research in the medical image registration community. Accurate registration of thoracic CT is both extremely useful in clinical terms and exceptionally challenging due to the elastic nature of lung tissue deformations. The goal of the EMPIRE10 challenge is to provide a platform for in-depth evaluation and fair comparison of available registration algorithms for this application. There are a number of benefits to comparing algorithms in this way:
- All algorithms will be applied to exactly the same set of data.
- Any algorithm parameters or settings will be chosen by those familiar with the algorithm and expert in its configuration.
- The resulting registrations will be independently evaluated using the same criteria for all participants.
How does it work?
The EMPIRE10 challenge was launched in early 2010 with an initial set of 20 scan pairs to be registered by participants in their own facility. This was followed in September by a workshop at the MICCAI 2010 conference where participants registered a further 10 scan pairs live within a 3 hour timeframe. This process and the results obtained are described in detail in Murphy et al., "Evaluation of registration methods on thoracic CT: the EMPIRE10 challenge.", IEEE Trans Med Imaging. 2011 Nov;30(11):1901-20. Please cite this publication if you wish to reference the EMPIRE10 challenge. From this point forward all participants will be judged based on the full set of 30 scan pairs. New participants and new submissions are always welcome - in this way we hope that the EMPIRE10 website will continue to reflect the state of the art in registration of pulmonary CT images.
On this website, teams can download a set of 30 scan pairs to be registered. For each registration they should produce 3 deformation field (displacement) images, one for each orthogonal direction (X,Y,Z) . These images should be submitted. Each set of results should be accompanied by a description of the algorithm in pdf format.
Who is organising the challenge?
EMPIRE10 was organised by the team members listed below:
- Keelin Murphy, University Medical Center Utrecht, The Netherlands
- Bram van Ginneken, Radboud University Nijmegen Medical Centre, The Netherlands
- Joseph M. Reinhardt, Department of Biomedical Engineering, The University of Iowa
- Sven Kabus, Philips Research, Hamburg, Germany
- Kai Ding, Department of Biomedical Engineering, The University of Iowa
- Xiang Deng, Siemens, Beijng, China
Josien Pluim, University Medical Center Utrecht, The Netherlands