Hi Alan,

We are currently in need to change the size and color of the arrows in the group difference visualization. I know you have told me that the size is fixed in ShapeworksStudio. Can you let me know if there are possibly work around methods there?

For example, will it generate larger arrows if the spacing value in the original nrrd data is increased? I think the shape modeling will need to be run again on the modified data and the change of the scale need to be tracked to project it back to the millimeter space which would not very convenient. Thus your insights will be very helpful.

For the color, I can change it to red/blue or others. It might be great already for what we want but I am still curious if there are more alternatives.

Thanks and regards!

Wenzheng

Thanks for your help, Alan! I find that configuring the scale only works when “Uniform scale” is checked. I didn’t have it checked previously and I thought it is not configurable. Now it looks good.

Thanks again!

Wenzheng

Hi Wenzheng,

1) Previous versions of Studio scaled the difference vectors based on the glyph size slider, but there was feedback that users wanted the vectors to show the actual length/magnitude.  We will have to add some option to switch between the two modes.

2) The colormaps are very configurable.  You can choose in preferences and you can adjust the lower and upper bounds at the bottom.

3) I’d suggest using Normals and possibly Geodesics.

Thanks,

Alan

Hi Alan,

Thanks a lot for your help! This feature is great and I have got some nice looking visualizations (shape mean and difference shown below). Can I ask a few more questions? Just curious about some finer adjustments.

1. Is there a way to get bigger arrows in the group difference visualization?

The arrows seem small and hard to see now. Perhaps this is just a result of too few particles.

2. Is there a way to adjust the color mapping in the group difference visualization?

It seem 0 is mapped to yellow in the current mapping. I think in our case it would be preferable to map it to black.

3. What could improve the optimization on very thin shapes?

It has some mismatching issues when the particle number is increased from 64 to 128 (shown below, figures from left to right show the deformation of the first principal component, mismatched areas are circled in orange).

Here are the original segmentation, groom parameters, groomed segmentation and optimization parameters:

Thank you!

Wenzheng

Hi Wenzheng,

Yes, this functionality is available in Studio, for example:

http://sciinstitute.github.io/ShapeWorks/latest/studio/studio-analyze.html#group

If you click the 'mean' for each group, you can export the reconstructed mesh from the File->Export menu.

Let me know if you have further questions.

Thanks,

Alan

On 8/1/22, 5:15 PM, "shapeworks-users-request@sci.utah.edu on behalf of wenzheng tao" <shapeworks-users-request@sci.utah.edu on behalf of wztao@cs.utah.edu> wrote:

    Hello,

    I am using Shapeworks to do shape analysis on some soft tissues and I

    would like to visualize the group difference in a way similar to the

    attached "norm_vec.png". It is a figure from [1], just as an example. I

    am wondering if Shapeworks has built in functions that can help me

    achieve the goals and I think there are mainly 2 questions.

    1. Is there a way to get a nice mean mesh for a subgroup?

    I assume it needs these steps: export the mean mesh from Shapeworks

    Studio, average the particles of the subjects in a selected group, warp

    the mean mesh to match the average particles. Does Shapeworks have these

    functions (I can do the averaging in Python)?

    2. Can shapeworks visualize normal vectors with colored arrows?

    I'd like to get a similar figure like the attached "norm_vec.png".

    Thanks for your help!

    Wenzhneg

    Ref:

    [1] Cates, Joshua, Lisa Nevell, Suresh I. Prajapati, Laura D. Nelon,

    Jerry Y. Chang, Matthew E. Randolph, Bernard Wood, Charles Keller, and

    Ross T. Whitaker. "Shape analysis of the basioccipital bone in

    Pax7-deficient mice." Scientific reports 7, no. 1 (2017): 1-10.