How Deep Learning Can Democratize Animation And VFX Posted Jun 14, 2020
It goes without saying that Artificial Intelligence (AI) is casting waves of disruption through a variety of industries and sectors. AI helps businesses to complete manual and repetitive tasks, and push the boundaries of what is possible without the need for counting on a large team.
One such industry that has major potential for AI is the online gaming sector. The gaming industry has been long at the vanguard of analyzing player behaviour. Using gambling techniques and other data collection methods, online gaming employs sophisticated systems aimed at understanding why players choose specific games, stop playing, or switch games.
Also, deep learning (a subset of machine learning in AI) can improve processes like animating characters and objects within the game, which would allow to save funds and make the process more accessible.
Read on to know how deep learning can democratize animation and VFX.
World’s first giant leap towards 3D programming
Back in the early 1950s, a young boy named Ed nursed a wild ambition to become a feature film animator. He loved Disney’s Pinocchio and Peter Pan and could not imagine a better way of leading life. But by the time he reached college, he realized that he lacked the necessary drawing skills.
Both Pinocchio and Peter Pan were hand-sketched, two-dimensional movies. This meant that their artists had to work hard and long to sketch each frame. They drew each character posing slightly different from its previous scene, and then some more, and more.
As the boy lacked artistic talent, he turned towards two areas he did have skills in - physics and computer science. Using these skills, he decided to make the first computer-generated animated film.
Eventually, he created a model of his left hand that he digitized in a self-written three-dimensional animation program for a 1976 film, Futureworld. The young boy’s computer-animated hand is considered the world’s giant leap towards 3D programming.
Intelligent animation
The shift from 2D to 3D was path-breaking. With 3D, artists could add a lot more details to their characters and treat them as physical objects. And instead of an artist sketching multiple characters in different angles, they could move the camera around the model seamlessly.
Recent advances in deep learning and AI have made it easier for artists to automate the laborious processes. It’s now possible to automate and speed up rigging, for example, by developing a basic skeleton for a character. Also, deep learning is used to present suggestive rigs for other characters.
Creating new characters is one of the most challenging jobs for an animator. Neural network technology can assist animators in doing that. By using the existing set of characters, neural networks can come up with new samples by learning from them.
Similarly, deep learning and AI can help in animating the character. It allows animators to transform what they’re creating in the style of someone whose work they admire.
Virtual scene creation is the next area AI can explore. The neural network can learn from existing images and synthesize new ones just like an artist would draw something.
