Title:	AUTOMATICALLY LEARNING AN INTUITIVE ANIMATION INTERFACE FROM A COLLECTION OF HUMAN MOTION CLIPS
Author(s):	Marcel Lüdi, Martin Guay, Brian McWilliams and Robert W. Sumner
ISBN:	978-989-8533-66-1
Editors:	Yingcai Xiao and Ajith P. Abraham
Year:	2017
Edition:	Single
Keywords:	Intuitive interface, character animation, motion manifold
Type:	Full Paper
First Page:	21
Last Page:	29
Language:	English
Cover:
Full Contents:	click to dowload
Paper Abstract:	In this paper, we automatically learn interpretable low dimensional generative representations of human walking motions using a variational autoencoder. By modeling the latent space of an autoencoder as a low dimensional multi-variate gaussian distribution, we can optimize for an encoding that produces disentangled, independent components which explain most of the variation in the data. The latent variables our model learns are intuitive to humans and can be directly manipulated in a graphical user interface (GUI) via sliders, to generate new walking motions in real time.
	Go Back