Hey /3/,I'm designing and presenting a medical research paper in a few months, and I want to build some 3D models out of patient CT/MRI data.I can convert the Dicom files (archive of images with positioning data) to various STL models but I want some way of cleaning up and assembling the pieces. I also want to make a majority of the model transparent, then isolate and highlight the spine and blood vessel information from the MRI.
>>565989this will by no means easy to do, you could have a look at something called "photogrammetry" (scanning photos to make badly made 3D models) then ones you have 3D data "retopology" (converting the badly made models into good ones) all the other things you want to do should be relatively simple.
>>565995no photoscan or stitching required, its a full body MRI which has all of the information needed to generate a dermal 'shell' and the internal structures I want. What I want is the best tool for cleaning up and assembling the 3D pieces. I intend to manually smooth and isolate the parts.I should also mention Id like some ability to animate in an info-graphic style so the m
>>566003you could do that in almost anything, if you don't own a software use blender
>>566003>What I want is the best tool for cleaning up and assembling the 3D pieces. I intend to manually smooth and isolate the parts.Blender and a lot of time. I can't think of any specific program that would be better for removing junk data from your scan. Once you get enough of the rough stuff out of it, you can apply smoothing parts of the mesh.
>>565989Don't listen to the other posters OP, it's pretty simple to do. Just grab a copy of Invesalius 3.0, import the dicom files and generate a surface based on the threshold you need. It's Open SourceFor cleanup you can try Meshlab, though you may not need it if the scan is good.
>>566084what this anon said sounds the best
>>566084w-what are you guys up to
>>566086Extracting geometry from medical data? It's hard to find good CT/MRI datasets though, people don't just share em the way they share models. I guess it's because there's usually patient info in the metadata.