Menu Close

This AI Framework Could Turn Your Sketches into 3D Objects

In the world of video editing and motion graphics, there’s usually a major skill jump required for moving between 2D editing and 3D compositions. There are a lot of reasons for this as the workflows become more technical and labor-intensive in a lot of ways.

However, in today’s age of AI, these jumps are becoming smaller and more manageable thanks to the wonderful art of machine learning and AI modeling. Case in point, a new sketch-to-3D-object AI tool has been introduced called to help you “Doodle Your 3D”.

In a paper by Hmrishav Bandyopadhyay, Subhadeep Koley, Ayan Das, Aneeshan Sain,
Pinaki Nath Chowdhury, Tao Xiang, Ayan Kumar Bhunia, and Yi-Zhe Song for the University of Surrey, the team introduced a framework that allows creatives to sketch a design and transform it into a 3D model.

Let’s take a look at this new AI modeling and alignment framework and explore what this could potentially mean for the future of 3D object design and motion graphics for filmmakers and content creators alike.


Doodle Your 3D

u200bAI 3D Object Shape Generation

In an effort to “democratize 3D content creation,” a new novel part-level modeling and alignment framework has been introduced on GitHub that aims to enable the precise generation of 3D shapes from abstract sketches. The goal of this project is to help content creators overcome any self-imposed limitations tied to drawing skills.

The paper gives some pretty abstract but technical breakdowns into what’s going on here, but for those of us who aren’t very AI-literate, it’s a bit hard to understand. The broad strokes, though, are that this new part-level modeling and alignment framework can facilitate abstraction modeling and cross-modal correspondence to eliminate the need for a data set pairing human sketches and 3D shapes.

The end results might be all you care about, though, and those are quite impressive indeed, as you can see some examples of this shape generation in action in these samples below.

3D Object Design and Modeling

Looking backward though at traditional digital 3D object design and modeling, it does seem like the industry is in need of a major innovation. And, truth be told, this isn’t the first 3D AI modeling tool out there. We’ve covered several other 3D AI modeling tools in the past as well.

However, as with all things AI, the greatest benefit that artificial intelligence and machine learning can offer us everyday content creators and video editors is simply the brute force help to handle a lot of the labor-intensive tasks that are put in front of us.

Many talented 3D designers and artists could easily take a sketch and turn it into a 3D model. Depending on their skill level and tools and their disposal, though, it would take anywhere from a few hours to a few days. With AI, though, this process could be sped up significantly, which would be quite beneficial.

The Future of 3D Content Creation

It’s interesting to see that this new AI is aimed so specifically at content creators, though. There are a host of other areas in which sketch-to-3D-object AI could be useful. However, while other industries like industrial design and home planning could benefit, content creation might simply be the most open-ended.

With AI tearing through the film and video world at a rapid pace now, it seems more likely than not that we’re going to see some of our first fully AI-generated projects launched here soon. Now whether those are full movies or just weird streaming content is still to be seen. But with other new content creation avenues like game design and mixed-used VR projects on the rise, the possibilities are starting to feel endless.

We’ll see where this new project ends up once it’s further developed. But, regardless of how you feel about it, this AI technology will only continue to expand in this space.

Author: Jourdan Aldredge
This article comes from No Film School and can be read on the original site.

Related Posts