alicevision / Meshroom

3D Reconstruction Software
http://alicevision.org
Other
10.83k stars 1.06k forks source link

Creating a 3D image from CT scans #2426

Open Sunny-UK opened 3 weeks ago

Sunny-UK commented 3 weeks ago

I would like to create a 3D image from a series of CT scans I had taken of an injury, however I cannot seem to find the right settings on Meshroom to make it work. I have a series of 30 photos, taken at 12° increments.
The images are only 906x898 but that is how I received them from the hospital files.

Does anyone have any suggestions to help me please?

Screenshots image

**Desktop Meshroom Version 2023.3.0 Windows 10 Python 3.7.4

msanta commented 3 weeks ago

I think the lighting and smooth shading used in these rendered images is going to make it impossible to reconstruct the 3D model.

fabiencastan commented 2 weeks ago

You probably need to find the focal/sensor size to get it working (the default value is 45° FOV and keeps the focal fully free). If you have a correct initial value, it could help a lot.

Sunny-UK commented 2 weeks ago

I think the lighting and smooth shading used in these rendered images is going to make it impossible to reconstruct the 3D model.

That's what I feared. On the one hand, you'd think having images equally spaced apart with no clutter would be easy to stitch together, but it feels like it needs more substance to work!

fabiencastan commented 2 weeks ago

Could you display the features detected? image

netron667590 commented 3 days ago

with absence of background information to anchor the object orientation it will be really hard, also for CT scan they should have raw data to generate the 3d model by itself.