**Use of this dataset must cite the following publications:**
[1] EL Brewer, LW Clements, JA Collins, DJ Doss, JS Heiselman, MI Miga, CD Pavas, and EH Wisdom III. "The Image-to-Physical Liver Registration Sparse Data Challenge." In Medical Imaging 2019: Image-Guided Procedures, Robotic Interventions, and Modeling, vol. 10951, pp. 364-370. SPIE, 2019.
[2] Collins, Jarrod A., Jared A. Weis, Jon S. Heiselman, Logan W. Clements, Amber L. Simpson, William R. Jarnagin, and Michael I. Miga. "Improving registration robustness for image-guided liver surgery in a novel human-to-phantom data framework." IEEE transactions on medical imaging 36, no. 7 (2017): 1502-1510.
[3]
**Data Files for Challenge**
Preoperative 3D Binary Liver Mask (LiverVolume.mhd, LiverVolume.zraw)
- LiverVolume.mhd
- LiverVolume.zraw
Preoperative 3D Liver Mesh (LiverVolume.nod, LiverVolume.elm, LiverVolume.bel)
- LiverVolume.nod
- LiverVolume.elm
- LiverVolume.bel
Intraoperative Sparse Data Patterns of Liver Surface
- DataSets.zip
Trial Data for Development
In the .zip file below, we have provided a subset of target information for 4 of the 112 challenge cases that were publicly available during the challenge period and can be used to guide development without revealing the full set of blinded target information for algorithm validation.
- AdditionalDataForParticipants.zip
The files are (LiverTargetsIncomplete.xyz, Set044_TargetsIncomplete.xyz, Set057_TargetsIncomplete.xyz, Set067_TargetsIncomplete.xyz, Set084_TargetsIncomplete.xyz). The first file is the location of a subset of targets within the 3D mesh provided (LiverVolume.nod). The next 4 files represent the new coordinate of the target after deformation. The deformation driving sparse data sets are from challenge sets 44, 57, 67, and 84 respectively.
**Results Submission**
*Full submission:*
A full data submission represents a compressed zip file (use standard zip archive in MSWindows environment) that contains one result file per sparse challenge data set for a total of 112 results files to be used for submission. It is critical that each result file have the following naming convention (name is within the single quote, case sensitive as well): ‘ResultsSet001.xyz’, ‘ResultsSet002.xyz’, ‘ResultsSet003.xyz’, etc… Each of these would correlate with the first, second, and third sparse data set driven results, and so on, respectively, and would continue until ‘ResultsSet112.xyz’ for a complete submission. All 112 files should be ascii file format and should be stored in one .zip file (MS Windows type). Each file will consist of 7 columns consisting of: an index, x nodal coordinate, y nodal coordinate, z nodal coordinate, displacement of nodal coordinate in x direction, displacement of nodal coordinate in y direction, displacement of nodal coordinate in z direction. We should note that the first 4 columns will be identical to the 4 columns in the provided mesh node file provided above. Once successfully submitted, results will be automatically processed, and the results will be posted on the Full Submission Dash Board shortly. Formal entry is only associated with a full submission. Partial entry information will be considered a submission, however, any reporting in a later publication will be designated as a partials submission.
*Partial submission:*
With a partial submission, the notes about the compressed file in the previous section are the same. The primary difference in a partial submission is that only a subset of the 112 results files is submitted in the .zip file. The most important aspect is to ensure that the number reference within ‘ResultsSet###.xyz’ matches the corresponding number as relayed in the sparse data challenge sets that are utilized to drive the result. Once .zip file containing some results is successfully created, results can be automatically processed with the provided analysis script.
**Previously Blinded Validation Data**
Previously blinded validation data and analysis code will be released upon publication of manuscript currently in review.
