realtimeDeeds

Code

Example code for TBME Publication "Model-based Sparse-to-dense Image Registration for Realtime Respiratory Motion Estimation in Image-guided Interventions".

Datasets

Datasets used in the experments:

• 2D MRI dataset [1] used in the code can be found under: https://zenodo.org/record/55345#.WvLINd8xBhF

  (additional files for example code can be downloaded here)

• 3D MRI dataset [2] and 3D US dataset [3] can be found under:

  [2] http://www.vision.ee.ethz.ch/en/datasets/ (4D MRI lung data)

  [3] http://clust.ethz.ch/data.html

[1]: CF Baumgartner, C Kolbitsch, JR McClelland, D Rueckert, AP King, Autoadaptive motion modelling for MR-based respiratory motion estimation, Medical Image Analysis (2016), http://dx.doi.org/10.1016/j.media.2016.06.005

[2]: Boye, D. et al. - Population based modeling of respiratory lung motion and prediction from partial information - Proc. SPIE 8669, Medical Imaging 2013: Image Processing, 86690U (March 13, 2013); doi:10.1117/12.2007076.

Landmarks

Manually selected landmarks for 2D/3D MRI and 3D US dataset are available as .txt (for 2D MRI and 3D US) and .m (for 3D MRI) files.

• 2D MRI: each txt-file contains landmark coordinates for one frame.
• 3D MRI: by executing MATLAB script file, 3 matrices (refLM, LM01px/LM04px, frames) are generated.
• 3D US: each txt-file contains landmark coordinates for all frames with frame numbers in the first column.