Create a Texture Generator with a GAN

Generative Adversarial Networks are a special type of Neural Network that can learn the probability distribution of a dataset. Wait, what?! Imagine you want to create a “procedural” marble texture generator with a Neural Network. You have a set of images you want to use as examples, those are the …

Implement a Substance like Normal Map Generator with a Convolutional Network

In this article, you’ll learn how to train a convolutional neural network to generate normal maps from color images. Convolutional neural networks are great at dealing with images, as well as other types of structured data. If you want to understand how they work, please read this other article first. …

Compress and Denoise MoCap with Autoencoders

In this article, you’ll learn how to train an autoencoding Neural Network to compress and denoise motion capture data and display it inside Maya Autoencoders are at the heart of some raytracer denoising and image upscaling (aka. super-resolution) technologies. Despite the pompous name, an autoencoder is just a Neural Network …

Interview: Dmytro Korolov on the results of SIGGRAPH’s AI tools for VFX panel

In this latest SIGGRAPH, Deep Learning was a hot topic. Many papers relied on it, some sessions were all about it. Dmytro Korolov, an experienced Pipeline Technical Director working at MPC, moderated a panel on the topic alongside Doug Roble (Digital Domain), Rob Pieke (MPC), Renaldas Zioma (Unity), Jeff Kember …

SIGGRAPH Series – Part 03 – ‘’DeepMimic: Example-Guided Deep Reinforcement Learning of Physics-Based Character Skills”

This is the third installment of the 3DeepLearner 2018 SIGGRAPH Series where I cover the most interesting deep learning papers (according to our audience) in a short video format. This video covers DeepMimic: Example-Guided Deep Reinforcement Learning of Physics-Based Character Skills by Peng et al. Link to the paper: https://xbpeng.github.io/projects/DeepMimic/index.html …

Teaching a Bop Bag to Stand Up with Simplified DeepMimic

DeepMimic relies on reinforcement learning to teach skills to ragdolls. It’s time for you to experiment with a simplified version of it in Maya. In previous posts, we have discussed the approximation of character animation with Neural Networks. This approximation was kinematic, that is, it was not concerned with the …

SIGGRAPH Series – Part 02 – ‘’Mode-Adaptive Neural Networks for Quadruped Motion Control”

The second installment in the SIGGRAPH series is here. The ‘Mode-Adaptive Neural Networks for Quadruped Motion Control‘ by Zhang, Stark, et al. was the most voted paper in our pool. So I think you’ll like this. I’ve also created a guide to the authors’ GitHub repository, check it out! Links …