Hi， my name is qinghong shao, my major is computer science, and my research direction is 3D printing. My major of my undergraduate is software engineering which give a solid foundation to design the software or the algorithms for the 3D printer. I want to know more knowledge about 3D printing, especially I want find or build a bridge between the AI(artificial intelligence), ML(machine learning), 3D Reconstruction and the area of 3D printing. And I want to make the 3D printing technology more accessible for kids and the people which is irrelevent to STEM major or work. If the staff of auto mechanic could use the 3D printer to print some metal part of the car and use it on the car, they don’t need order these part online which will take many days to receive this parcel, I think it is what I want to see in future.
https://github.com/nvlabs/instant-ngp
The project I choose is published in 2022 by NVIDIA’s lab when i was in my third year of university. When i see this algorithm i guess i fall in love with it because its code is efficient although isn’t simple, the result is also impressivel. With the mobile phone , user can shoot around the object, and use the algorithm to render for a few minutes to get the corresponding 3D object model, and it works. There are still many limitation for this algorithm.
The researchers in Lab of NVIDIA they are focus on the research of computer vision and image processing Jonathan StephensJonathan. They have finish many projects about the Engineering, Construction, 3D rendering and so on. I really appreciate the all of the work they have done.
Why this algorithm attracts me, maybe it will be complex and obscure to explain. In simple terms, these algorithm didn’t use the traditional explicit representation, but the implicit representation. The result show this implicit representation, the NeRF(neural radiance field) is efficient, it use functions to express 3d objects, reduce storage space and computing costs.
Finally, what I want to mention is that many lab they focus on the improvement of the 3D reconstruction algorithms, but I choose the other road, develop the 3D printer so that it could print unimaginative ideas, maybe generated by AI, or it just occurred you and you want print it at now. I believe there are still many works left to deal with in the direction.
Hello Quinghong,
I think the project you chose was very interesting. Although I am not an algorithms type of guy, I can still appreciate a complex and efficient algorithm. I am also interested in learning about 3D printing with AI and machine learning. I completely agree that there are so many directions we can take with 3D printing and that there are many projects that we haven’t even begun thinking about.
Hi Qinghong,
First reading your response I was a little surprised by the seemingly sudden direction change from a focus on Software Engineering to 3D Printing, but your motivations to make 3D printing more accessible to inexperienced users ties those two study areas quite nicely. Its very inspiring, and I hope that your research focus puts you closer to these goals.
Also, I’ve never heard of your inspiring project before, but at first glance, I might fall in love with it as well! It also seems very close to the path you see for 3D modeling moving forward. Your explanation of why the project attracts you does seem very high-level, enough to grab my attention and convince me to look into it (particularly NeRFs) more. I found this article on it (https://www.matthewtancik.com/nerf). While I don’t really have the math background to understand all of it, I think its going to be an interesting read.
Thanks for sharing!
Wow, thanks for sharing this project! The way they use functions to express 3D objects is pretty impressive! I wouldn’t say I fully understand it, but I love finding implicit ways to do things more efficiently!
And of course the ramifications of this kind of technology are incredible. To be able to scan a real-world object and create a 3D model of it really makes a lot of processes much faster!