Python 3D Printing Guide: Make STL Files From Masks

This video shows how to take a 3D numpy mask of 1's and 0's and turn it into a corresponding STL file that can be 3D printed or viewed using a 3D viewer on your computer.

NumPy Meshgrid Tutorial:

Join the Discord Server:

[cooperbarry]

Thank you for the great video! It was instrumental for me to 3D print antenna and array beam patterns.

I found a way to smooth the stl instead of getting a "blocky" shape without having to resort to smaller dimensional spacing. Take the 3D mask array, cast it to single, and do a Gaussian blur (e.g., The larger the sigma value in the Gaussian blur, the smoother the surface will be but at cost of losing small and sharp features. (The Gaussian blur is low-pass filtering the image.) The marching cubes algorithm will then interpret values on [0, 1] as a smooth surface, whereas if it were a bool mask on {0, 1}, the marching cubes algorithm makes a blocky surface with sharp, 90-deg corners between pixels.

[tiddlywinks]

Very nice! I use .stl files daily from material imaging, so it's cool to be able to understand the process now.

[fedyasemenov]

Dude you are legit genius! Such a complex topic explained with such detail and clarity. The sign of a true master. Understanding this made me feel good about myself. And you are the one who made me understand it.

Thank you so much 🙏. Please keeps it up!

[varunahlawat]

I'm a beginner and physics enthusiast. This made me a fan

[Fuxx]

i have a friend, who got a small fracture in his elbow. He printed out the scans of his elbow similarly to this method here. First time I actually found 3D printing kinda useful.

[rosealsendi]

Thanks for the great video! Really impressed with your python & physics content.
I have a recommendation for a future python video; I think it'd be great if you make a plotly tutorial, because it's a very useful library and you've used it before in your physics videos. Also because your library tutorials include many realistic and useful examples especially for physicists.

[ogiycsf]

I was quite confused when going through Numpy-stl documentation... So a big thank you for this video!
For those who wonder how to smooth the 3d object, modify as follows:
x=y=np.linspace(-1, 1, 100)
z=np.linspace(0, 0.5, 100)
to
x=y=np.linspace(-1, 1, 500)
z=np.linspace(0, 0.5, 500) it's not rocket science but I hope it helps!

[TheBarryLight]

Now I know what I will do on the next weekend :D

[Jester]

Very cool topic! I’m impressed by how you present new (to me) topics and techniques in a clear and digestible fashion.

Are you interested in making videos on finite element analysis, and do you think this video’s topic would be applicable to FEA?

[igorbrenno]

I love you, brow. Thanksss...so much!

[noahmcallister]

Great video! Have you heard of the program called MeshLab? I'm fairly certain it has all the capabilities of what you are doing here as a GUI, as well as Python bindings. I've used the Python API for procedurally generating and decimating meshes for a research project; it was super easy to use. Thanks again!

[nicholasmetsavaht]

Great video! Can I have more than one function to define the object, like from 0 <= z <= 4 use a certain function and from 4 < z <=10 another that coincides with the last points from the previous.

[mrmaple]

Thanks for this. I've done marching cubes and 3d printing in the past, but never with numpy! Great stuff. Do you ever use VTK / ITK for visualization and manipulation, or are they less relevant these days?

[MinerMovie]

This video was great, helped me a lot with a project I'm working on. One thing, it would be nice to have the code in the video somewhere

[russelldicken]

you've not come across democratiz3D?