Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Issue with Softbody Simulation for Vascular Modeling
#1
Hi,
I am writing to ask for help.I'm currently working on simulating a vascular wall using a model of blood vessels with Blender's softbody physics. However, I'm facing some problem with the particle not conforming well to the vascular walls.
When generating the blueprint for the entire vascular model, the particles do not adhere properly to the vascular walls. Even when I import individual sections of the model into the softbody blueprint, the issue persists.
I tried dividing the model using Blender, and interestingly, using the bottom four vessels produces better-fitting particles. However, as I increase the length of the divided vessels, the simulation quality deteriorates(the number of particles reduced).
On further inspection, I noticed that when I isolate a portion of the top vessels and attempt to generate particles, they consistently fail to conform to the vascular walls.

I've attached images illustrating these issues.
[color=var(--tw-prose-bold)]My question is: How can I modify either the model or the settings during blueprint generation to address these issues? [/color]
Thanks in advice for any help.
Reply
#2
Hi,

Can't see any images attached to your post. Either paste links to your images, or if attaching them directly to the forum make sure to press the "AddAttachment" button after adding each image.

regards,
Reply
#3
                   
Reply
#4
Hi,

Using particles to simulate a mesh like this one is not a good idea at all. The reason is that ellipsoids are a terrible at approximating thin, long shapes, which this mesh is almost exclusively made of. For modeling vascular systems, cosserat rods are a much better approach, and tetrahedral meshes can do a good job too.

If you really want to use a particle-based engine for this, I'd suggest:

- Drop voxel sampling entirely. You'd require extremely small voxels to pick up such thin features.

- Subdivide your mesh such that edge loops (and hence, vertices) are evenly distributed throughout its surface, then use vertex sampling instead (that is, set "Surface sampling" mode to "Vertices" instead of "Voxels")


Quote:I tried dividing the model using Blender, and interestingly, using the bottom four vessels produces better-fitting particles. However, as I increase the length of the divided vessels, the simulation quality deteriorates(the number of particles reduced).

The reason why adding longer vessels seemingly reduces the amount of particles used is because the bounding box becomes larger, and since your voxel resolution (which determines the amount of voxels used for the longest axis of the bounding box) is kept fixed, the amount of voxels used to sample the mesh is the same which means larger voxels must be used. Note you can enable "Voxels" in the Render modes tab at the bottom of the blueprint editor to visualize voxelization.

All these issues can be worked around by using vertex sampling instead of voxel sampling, as mentioned above. Still, ellipsoidal particles can only do so much when faced with such thin and long structures.

kind regards,
Reply
#5
(10-01-2024, 08:25 AM)josemendez Wrote: Hi,

Using particles to simulate a mesh like this one is not a good idea at all. The reason is that ellipsoids are a terrible at approximating thin, long shapes, which this mesh is almost exclusively made of. For modeling vascular systems, cosserat rods are a much better approach, and tetrahedral meshes can do a good job too.

If you really want to use a particle-based engine for this, I'd suggest:

- Drop voxel sampling entirely. You'd require extremely small voxels to pick up such thin features.

- Subdivide your mesh such that edge loops (and hence, vertices) are evenly distributed throughout its surface, then use vertex sampling instead (that is, set "Surface sampling" mode to "Vertices" instead of "Voxels")



The reason why adding longer vessels seemingly reduces the amount of particles used is because the bounding box becomes larger, and since your voxel resolution (which determines the amount of voxels used for the longest axis of the bounding box) is kept fixed, the amount of voxels used to sample the mesh is the same which means larger voxels must be used. Note you can enable "Voxels" in the Render modes tab at the bottom of the blueprint editor to visualize voxelization.

All these issues can be worked around by using vertex sampling instead of voxel sampling, as mentioned above. Still, ellipsoidal particles can only do so much when faced with such thin and long structures.

kind regards,
Thanks for your prompt and detailed response. I will try based on your suggestions.
With appreciation!
Reply