Phase-Functioned Neural Networks for Character Control
We present a real-time character control mechanism using a novel neural network architecture called a Phase-Functioned Neural Network. In this network structure, the weights are computed via a cyclic function which uses the phase as an input. Along with the phase, our system takes as input user controls, the previous state of the character, the geometry of the scene, and automatically produces high quality motions that achieve the desired user control. The entire network is trained in an end-to-end fashion on a large dataset composed of locomotion such as walking, running, jumping, and climbing movements fitted into virtual environments. Our system can therefore automatically produce motions where the character adapts to different geometric environments such as walking and running over rough terrain, climbing over large rocks, jumping over obstacles, and crouching under low ceilings. Our network architecture produces higher quality results than time-series autoregressive models such as LSTMs as it deals explicitly with the latent variable of motion relating to the phase. Once trained, our system is also extremely fast and compact, requiring only milliseconds of execution time and a few megabytes of memory, even when trained on gigabytes of motion data. Our work is most appropriate for controlling characters in interactive scenes such as computer games and virtual reality systems.
Видео Phase-Functioned Neural Networks for Character Control канала Yoshiboy2
Видео Phase-Functioned Neural Networks for Character Control канала Yoshiboy2
Показать
Комментарии отсутствуют
Информация о видео
Другие видео канала
![SIGGRAPH 2020 | Learned Motion Matching](https://i.ytimg.com/vi/16CHDQK4W5k/default.jpg)
![Flexible Muscle-Based Locomotion for Bipedal Creatures](https://i.ytimg.com/vi/pgaEE27nsQw/default.jpg)
![Animation Bootcamp: An Indie Approach to Procedural Animation](https://i.ytimg.com/vi/LNidsMesxSE/default.jpg)
![IK Rig: Procedural Pose Animation](https://i.ytimg.com/vi/KLjTU0yKS00/default.jpg)
![AI Cars learn to drive - UE4 Blueprints](https://i.ytimg.com/vi/yjrxokMq8Rk/default.jpg)
![A Deep Learning Framework for Character Motion Synthesis and Editing (SIGGRAPH 2016)](https://i.ytimg.com/vi/urf-AAIwNYk/default.jpg)
![Google DeepMind AI Does Parkour - Producing flexible behaviours in simulated environments](https://i.ytimg.com/vi/g59nSURxYgk/default.jpg)
![Animate a character using PROCEDURAL Animations ONLY! - Tutorial Unity3D - Procedural animation #1](https://i.ytimg.com/vi/AChwSWU4AaU/default.jpg)
![[SIGGRAPH Asia 2019] Neural State Machine for Character-Scene Interactions](https://i.ytimg.com/vi/7c6oQP1u2eQ/default.jpg)
![Motion matching for Unity (own implementation, link to github in description)](https://i.ytimg.com/vi/ihzNX3ASQx4/default.jpg)
![Google's DeepMind AI Just Taught Itself To Walk](https://i.ytimg.com/vi/gn4nRCC9TwQ/default.jpg)
![[SIGGRAPH 2018] Mode-Adaptive Neural Networks for Quadruped Motion Control](https://i.ytimg.com/vi/uFJvRYtjQ4c/default.jpg)
![Character Control with Neural Networks and Machine Learning](https://i.ytimg.com/vi/o-QLSjSSyVk/default.jpg)
![This AI Learned To Animate Humanoids🚶](https://i.ytimg.com/vi/cTqVhcrilrE/default.jpg)
![[Nucl.ai 2015] Motion Matching - The Road to Next Gen Animation](https://i.ytimg.com/vi/z_wpgHFSWss/default.jpg)
![Evolution of Neural Networks using Genetic Algorithm for a 3D car made in Unity](https://i.ytimg.com/vi/8V2sX9BhAW8/default.jpg)
![Learning Motion Manifolds with Convolutional Autoencoders](https://i.ytimg.com/vi/dLopOB6D9co/default.jpg)
![Citizencon 2017: Teaching a character how to walk on any terrain](https://i.ytimg.com/vi/PryJ3CpHcXQ/default.jpg)
![Procedural walk animation](https://i.ytimg.com/vi/U8jO8r9wWTM/default.jpg)
![Is a Realistic Water Bubble Simulation Possible?](https://i.ytimg.com/vi/O-52enqUSNw/default.jpg)