Hello folks,
I am working on a geometrical analysis project for some ~40 aortic models. I have achieved the following optimization with particles not perfectly splattered across however, they kind of are. I am currently trying to maximize the amount of information i can gather from this analysis. I am aware that I am able to extract the mesh of the optimized model; however, I’m interested in other information, such as the ability to obtain the mean , standard deviation, etc. I also wonder what the next steps are in the ShapeWorks process pipeline to obtain valuable information based on this analysis. How can I use the exported csv file and such info would be great.
Any comments or ideas are welcome and appreciated.
I will be attaching some figures for the results and the optimization parameters that I have used.
I also do struggle quite a bit with the optimization parameters and would love to get some explanations on what to choose based on what.
I am aware of the metrics section but I was wondering what else can I obtain from this lovely tool.
It depends what type of analysis you want to perform. Are you hoping to describe a population, or do you have groups that you’d like to characterize differences between (e.g. pathology vs control)?
Often times we export the PCA Component scores and then use R to test for significance of each mode of variation against clinical factors.
I don’t have any particular advice on the optimization parameters. This looks like challenging data as you have different morphologies that may not have corresponding features. That doesn’t mean you can’t analyze them with ShapeWorks, but it does pose some challenges.
I would like to thank you for taking the time to reply to the post. I really do appreciate the insight, thank you for the explanation. Within the frame work of this project i aim to carry some sort of pathology vs control. However, till then, I also should describe the population of these models, what would the best way to do so? What kind of data would I be able to extract?
Thank you once more for your efforts and thanks for creating ShapeWorks!
Export the mesh(es) from ShapeWorks: This will include your scalar data as well. If you want grouped data (given you have pathology vs control), go to the “Group” tab in the Analyze section, select the “Group set” name and export meshes for each group. If you have checked the box underneath saying “Show difference to mean”, these exported files will also contain the surface differences with respect to the mean surface, for each group.
Paraview is also an open-source software. Download and open the mesh(es) you exported from ShapeWorks. (Make sure you press the apply button)
Right click on the mesh, in “Add filter” option, choose “Convert to point cloud”.
From here onwards, you could either use Paraview itself to compute statistics and draw graphs, or if you prefer to use Python, just “Save data” as .csv.
One way to describe a population is by examining the modes of variation. In the PCA tab of analysis, you can see what the major modes of variation within your data are. You can also export via File->Export->PCA Montage an 3xN image showing these images from the same view. For example:
