On May 28, 2020, at 12:31 PM, AMY LORRAINE LENZ <amy.lenz@utah.edu> wrote:

Hi Shireen and ShapeWorks Team,

Thanks for the great updates in the monthly meeting.  Here are my requests and or bugs that I found when using and modifying the recent use cases in version 5.3.

1) When in Studio and visualizing the mode of variation results, I often export the mean and +/- 2SD surface meshes to perform post process visualization comparisons.  I found that using File->Export Surface Mesh, the .vtk file generated only contains the correspondence particle x,y,z locations for that mode of variation and not the reconstructed surface mesh itself.  The export gui says it is exporting as a .vtk, but then there is not a file extension that makes it recognizable as a .vtk file to common softwares.  So I just renamed it adding .vtk file extension but if you open that in a software such as CloudCompare, you'll just see the correspondence particles and not the mesh reconstruction.  I've attached two examples of the mean fibula file, with and without the .vtk extension.  Here is the same mean fibula surface reconstruction as exported by View2 in an old version that successfully exports with the .vtk file format as a mesh (Fibula_mean_smooth.vtk).  

2) Can you provide instructions on how to re-visualize the analyze.xml after the initial correspondence model has been executed?  For example, it may take a while to run the model, successful results are achieved  

3) Trying to build from Master on Ubuntu 18.10: Following build.md instructions, I got stuck trying to figure the correct cmake options.  This particular error, I never got past.  But when using the binaries from the latest release in the conda environment, it works just fine.  So your suggestion of providing binaries of the master, would eliminate this problem and not require a build.  This would enable testing of the new master branch features.  

4) A feature request.  Include statistical analyses in Studio for example: 1) Parallel analysis for containing significant PCA modes of variation.  2) Within these significant modes, PCA loading values compared between two groups, a student's T-test with Finner's adjustment for multiple comparisons would be great, and 3) Hotelling's T^2 test for determining significant shape differences between two groups.

Thanks so much!  Let me know if you have questions.

Best,

Amy

****************************************************

Amy L. Lenz, PhD

Research Instructor

Orthopaedic Research Laboratory

University of Utah Orthopaedics

590 Wakara Way, RM A118

****************************************************

<Fibula_mean><Fibula_mean.vtk><Fibula_mean_smooth.vtk>