Ortho Project : Scan (step 2)

The 3D scan will be a real challenge : the goal is to get the best possible scan of my jaw, because at the end it will impact the movement of my teeth. At this scale, millimeter counts.

At first, I tried to look for “fab lab” around Paris where I could eventually find a 3D scanner to use. Dead end, either the lab were definitively closed or there were too far from my house or work.

I found a company on internet, in Paris also, that realized 3D scan with professional scanner : the quote for my jaw was around 200€. No way ! I had to find something cheaper. (I was not even sure of the result at this time, and didn’t want to engage to many money)

Finally, parsing the infinite internet, I found solutions to do it myself. Why not, I really enjoy do things only by myself !
I discovered the photogrammetry : a technique that transforms multiples images of an object into a 3d object. Nice, that’s what I need !
First I tried mobile app on my phone, but the result was quite cheap : a smartphone cannot process so many data.
But then, I found Meshroom ! #holygraal

Meshroom is an open source photogrammetry software, almost usable by everyone.

The first tries I did were not very conclusive, but I look for advices to improve the scan, and some French guys on Youtube give me the perfect trick : dots !
As Meshroom will try to match different pictures from my model, I needed to help him with points that it can identify and recognize.
It works like a charm (after 2 hours of compute, with a hundred picture), and I get a very nice 3d model with only my smartphone as a camera and a computer !

Awesome model, with no disturbances

My 3D model is there, the first challenge is a success !

Next step, the manipulation of the 3D model.

Ortho Project : 3D printing (step 4)

That’s the easy part, with no real challenge, as I already own a 3D printer (gift from my previous colleagues, on my wife’s precious advice. I love you !)
I’ve got a DiscoEasy 200, from Dagoma (a french company) : it’s a nice 3D printer to begin with, itself partly printed.

With my 3D model ready, I “just” had to export it in STL format, in order to print it.
Yeah “just”.. I had forgot a basic universal rule !

A thing I completely forgot since the beginning, is that “size does matter” !

After the creation of my first aligner and its try for a few hours, I realized that something was wrong (as it was quite painful !) The 3D scan was one thing, but there was a challenge to modelize my jaw at the exact same size. My first 3D model was not precise enough, and had less than 1mm wrong in size.

So I had to resize it precisely, with 2 iterations of resizing-printing-trying. I measure the final 3D print a caliper on multiples points, and also try it with my upper jaw.

About the 3D printing, I used :

      • Slic3r, an open-source slicer for converting my 3D model into a printable file.
      • PLA material
      • 0,2 mm nozzle (for the 3D printer)
      • Infill 10% (a 3D printing parameter to have a solid core)

Each print is about 1 hour, and use around 8 gr of PLA.

Next step (and final ?), the aligner’s thermoforming !

Ortho Project : 3D edition (step 3)

Once I get the 3D model from Meshroom, the next challenge was to exploit this model : clean it up a little, isolate teeth and make them move.

For that, I use Blender, another great free and opensource software. However, in the discovering of Blender, I had the impression to work with Photoshop : it’s a powerful tool, but quick hard to event start for a beginner.
The learning was a bit long and I had to read some articles and watch multiples Youtube tutorials.

An important thing is to have a model at the right scale : Meshroom create a 3D model, but I had to scale it. For that, I use a pide-à coulisse to measure precisely the plaster mold and scale the 3D model on Blender.
(I mess up my first scaling, and the first aligner that I made was to small and I felt it on my teeth, it was hurting)
So to improve my scaling, I went back and forth between the 3D model that I 3D printed to re-measure and re-scale.

The useful functions I found were:

      • Ctrl+L : Select all linked vertices, really helpful to remove undesired independants objects
      • Intersect (Knife) tool : to have a plan base, I use this post’s answer
      • Mesh > Clean Up > Decimate Geometry : Reduce the number of vertices and faces on non-essential part of the model. I ended up with a model with 70k vertices (~14Mb)
      • Circle select (c) : To select part of the model to isolate a tooth. Combine with “Esc” to exit the circle tool, I could jungle with selection and moving into the model.
      • Ctrl+H > Hook to new object : Once a tooth selected, I could create a separate object in order to move it later
I also added a reinforcement at the back to strenthen the 3D printed model, and a scale to identify each print.

At this time, I had discussion about my project with 2 people that had aligners. I had learn some interesting informations:

      • The first aligner hurt a little bit more than the others
      • Teeth “naturally” move forward
      • In order to realign front teeth, sometimes I will have to create space by moving back the molars
      • Top and bottow jaw are “linked” : align one of them will have an impact on the other. So I decided to do aligners for the top too (and correct a minor thing)
      • The time between two aligners is around 2 weeks

After, I “just” had to do the animation, in order to have multiples steps to print.
I read here that teeth can move from 0,4 to 1,2/per month. Another website says 0,2mm per week (or 2 degre).
To be prudent, I choose to be under 0,2 mm per week. (or at least try)

For animation, I also had to learn how it works:

      • I > LocRot : Add a keyframe. One at the begining, one at the end (LocRot as LocationRotation)
      • G : move the object
      • R : rotate the object

At the end, the gum is deformed, but only the teeth count, as the aligner only need them. However, I’m a bit skeptical for the last move (front tooth), I may need to work more on it (and slow it down).

The model is clear, and the animation is fine, I solved the second challenge !

Next step, the aligner’s printing !