Scripts used for the conference proceeding "B.A. Besler, A.S. Michalski, N.D. Forkert, S.K. Boyd. “Automatic Full Femur Segmentation from Computed Tomography Datasets using an Atlas-Based Approach”. Computational Methods and Clinical Applications in Musculoskeletal Imaging: 5th International Workshop and Challenge, MSKI 2017, Held in Conjunction with MICCAI 2017. Quebec City, September 10th–14th, 2017."
.
├── COM
│ ├── helperScripts
│ └── imageProc
└── DOCCOM contains scripts used for this project. VTK 7 and simpleitk should be used for executing those scripts.
helperScripts contains scripts used for verification and checking.
imageProc contains scripts which did some image processing, such as thresholding or dilation. This also contains the Elastix files.
The Krcah segmentation technique is available online.
This project also contains some excellent datasets for testing against.
I was only able to build the project against ITK v3.10.2.
This can be checked out as a tag from the online git repository.
When building ITK, enable flags -std=c++0x –fpermissive and include ITK_REVIEW.
Elastix version 4.8 was used for thie project.
It can be attained from the Elastix Website.
For the selection criterion, first run the Affine.txt file with your data
Metrics can be grabbed using grep (grep -r -a "Final Metric: " *).
Sort the metrics by hand and run the best ranked metric with the BSpline.txt file.
Example data can be found in the Krcah repository. We are working on publishing a larger dataset containing the images from this study. This information will be updated when that happens. Please contact us if you would like access to the origin data.