Hi there! I’m looking to microscribe and 3D digitise anatomical specimens for an honours project, but struggling to make accurate 3D images in rhino from what I digitise. I’ve been using the curve point feature to outline my specimen and then multiple points inside, then patching to hopefully be able to fill it in. Unfortuantely the patches always are way too big for the object or miss out lots of the points. I’ve read patching works best with symmetrical objects so is that maybe why it isn’t working?
Is there an easier way to do this or a way that works better when microscribing complex 3D objects? Would love some help and guidance!
Photogrammetry will capture many more points than the Microscribe. There are some low cost, or even free photogrammetry applications that can produce a mesh based on good photos. There is a methodology to the picture taking, but you could get that info from reading online
The downside of photogrammetry is that is does not capture the actual size of the object, so it needs to be scaled. The scaling is calculated by measuring points on the resulting mesh and corresponding points on the object. For organic objects, and more accurate scaling, a ruler is usually placed in the photos for a reference. In your case, since you have the Microscribe, it could be a better tool than the ruler to reference the dimensions.
We have an ARTEC scanner which I have access to, so could definitely look into using that for more accurate modelling, but if there’s a way to solely use the microscribe I’d like to explore that first.
Theres quite a few papers that used the microscribe to digtise/measure accurate distances between anatomical landmarks, but they’re all very vauge on their actual 3D rendering/modelling process.
Regarding using the curves and points, I’m wondering if maybe because the anatomy isn’t symmetrical/even curves maybe I’m not doing enough curves or points for the porgram to understand it?
I won’t be back at the lab until the first week of the new year, so I’ll be able to post an example of the curves/points I have when I’m back! DO you reccommend more curves and points for more complex objects?
What type of anatomy, Cross sectional? Human, animal, insect? Depending on what you are doing it sounds like a daunting task.
If it is cross sectional anatomy I would scan it in 2d or photo it with a ruler. Then bring the picture into rhino and scale it to the photographed ruler and measure it there.
I would use the scanner or photogrammetry, digitizing can be error prone if you don’t know what you are doing, you really have to be set up to use it properly. You have to have a baseline and the subject and templates should have registration pins to be able to come back to your subject.
With photogrammetry or scanning you have the textures and everything ready to measure, you have to do clean up but it won’t be as hard as modeling the subject from points. Rhino’s Patch is not a robust command and if you’re really looking to model accurate anatomy be prepared for modeling headaches. You might be better off using sculpting programs like 3dcoat or zbrush even blender. David is doing boats and ships so his surfaces are of a certain type that are doable in Rhino. Anatomical subjects might not be suitable for this type of work flow depending on the complexity of the subject.
Epic games has a good free photogrammetry program and then there are also paid alternatives and you already have a scanner.
The object could require some fixturing or setup for the best capture, so its difficult to offer feedback on the process without knowing what the object is.
If you have access to an Eva or Leo, that will be a better option than the Microscribe or photogrammetry. You would have to learn to use the Artec Studio software to process the scanning for the best results before exporting a file to use in Rhino.
Human antatomy, aiming to look at nerves and musculature of the shoulder, but currently just trying to microscribe a scapula bone.
Some of the anatomical papers I’ve looked at have used a mixture of light scanning for the modelling and then digitising for the actual measurements/accuracy, so I might look in to doing that.
Yeh that makes sense. I’m currently trying to microscirbe a scapula bone, but the project will be looking at shoulder anatomy. Some papers I’ve read doing a similar thing used surgical plates to secure the specimen to microscribe, so I think doing that and maybe incoroporating using a scanner (we have a LEO and a SPIDER) I could maybe get accurate modelling alongside accurate measurements?
I don’t understand what purpose the Microscribe would serve if you have a Spider. I have a an old Microscribe and its accuracy is 0.009”. The Spider will give you better accuracy. You can get any point to point measurement you want from the mesh. Maybe I am missing something?
That might be the way to go but I think you can just measure from the scan and do away with measurements using the microscribe. I would use an accurate digital caliper with usb connection to check. You might want to get into Grasshopper too, you can automate and check numbers and measurements against your scans or slicing and making 3dprints. Shrink wrap command in V8 might help too to get a nurbs model from your scanned mesh.
The Microscribe accuracy is either +/-0.009” or +/- 0.002” depending on the model. The Spider accuracy is +/- 0.002”.
Using either the scanner or the digitizer may require repositioning the object to capture both “sides” for complete coverage. Aligning the 2 sets of scan data will be more accurate using the larger number of points the Spider will capture.
It may be possible to get the complete surface in one scanning session, assuming the fixturing method does not obscure the required details or impede the ability of the scanner / digitizer to move freely around the object.
Ideally though we want to use the microscribe in some way - as we already have quite a few projects going using the 3D scanners, so we want something with the microscribe as it isn’t being utilised at the moment. That’s why I was thinking maybe the most fesible way would be doing a mixture.
We have the 0.002 accuracy one. Ideally we want to use the microscribe in some way as it isn’t being utilised at the moment in the lab. Also in medical literature there’s a lot of potential as it hasn’t been as widely used as scanning due to the challenging properties and logistics of using it.
I think I’m leaning towards the idea of utilising both scanning and microscribe data