Learning Motion Manifolds with Convolutional Autoencoders

Daniel Holden; Jun Saito; Taku Komura; Thomas Joyce

doi:10.1145/2820903.2820918

Learning Motion Manifolds with Convolutional Autoencoders

Daniel Holden, Jun Saito, Taku Komura, Thomas Joyce

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract / Description of output

We present a technique for learning a manifold of human motion data using Convolutional Autoencoders. Our approach is capable of learning a manifold on the complete CMU database of human motion. This manifold can be treated as a prior probability distribution over human motion data, which has many applications in animation research, including projecting invalid or corrupt motion onto the manifold for removing error, computing similarity between motions using geodesic distance along the manifold, and interpolation of motion along the manifold for avoiding blending artefacts.

Original language	English
Title of host publication	Proceeding SA '15 SIGGRAPH Asia 2015 Technical Briefs
Place of Publication	New York, NY, USA
Publisher	ACM
Number of pages	4
ISBN (Print)	978-1-4503-3930-8
DOIs	https://doi.org/10.1145/2820903.2820918
Publication status	Published - 2015

Keywords / Materials (for Non-textual outputs)

animation, autoencoding, character animation, convolutional neural networks, deep neural networks, machine learning, manifold learning, motion data

Access to Document

10.1145/2820903.2820918

motioncnnAccepted author manuscript, 436 KB

http://doi.acm.org/10.1145/2820903.2820918

Cite this

@inproceedings{abdede47b0724c46819572583676add0,

title = "Learning Motion Manifolds with Convolutional Autoencoders",

abstract = "We present a technique for learning a manifold of human motion data using Convolutional Autoencoders. Our approach is capable of learning a manifold on the complete CMU database of human motion. This manifold can be treated as a prior probability distribution over human motion data, which has many applications in animation research, including projecting invalid or corrupt motion onto the manifold for removing error, computing similarity between motions using geodesic distance along the manifold, and interpolation of motion along the manifold for avoiding blending artefacts.",

keywords = "animation, autoencoding, character animation, convolutional neural networks, deep neural networks, machine learning, manifold learning, motion data",

author = "Daniel Holden and Jun Saito and Taku Komura and Thomas Joyce",

year = "2015",

doi = "10.1145/2820903.2820918",

language = "English",

isbn = "978-1-4503-3930-8",

booktitle = "Proceeding SA '15 SIGGRAPH Asia 2015 Technical Briefs",

publisher = "ACM",

}

TY - GEN

T1 - Learning Motion Manifolds with Convolutional Autoencoders

AU - Holden, Daniel

AU - Saito, Jun

AU - Komura, Taku

AU - Joyce, Thomas

PY - 2015

Y1 - 2015

N2 - We present a technique for learning a manifold of human motion data using Convolutional Autoencoders. Our approach is capable of learning a manifold on the complete CMU database of human motion. This manifold can be treated as a prior probability distribution over human motion data, which has many applications in animation research, including projecting invalid or corrupt motion onto the manifold for removing error, computing similarity between motions using geodesic distance along the manifold, and interpolation of motion along the manifold for avoiding blending artefacts.

AB - We present a technique for learning a manifold of human motion data using Convolutional Autoencoders. Our approach is capable of learning a manifold on the complete CMU database of human motion. This manifold can be treated as a prior probability distribution over human motion data, which has many applications in animation research, including projecting invalid or corrupt motion onto the manifold for removing error, computing similarity between motions using geodesic distance along the manifold, and interpolation of motion along the manifold for avoiding blending artefacts.

KW - animation, autoencoding, character animation, convolutional neural networks, deep neural networks, machine learning, manifold learning, motion data

U2 - 10.1145/2820903.2820918

DO - 10.1145/2820903.2820918

M3 - Conference contribution

SN - 978-1-4503-3930-8

BT - Proceeding SA '15 SIGGRAPH Asia 2015 Technical Briefs

PB - ACM

CY - New York, NY, USA

ER -

Learning Motion Manifolds with Convolutional Autoencoders

Abstract / Description of output

Keywords / Materials (for Non-textual outputs)

Access to Document

Fingerprint

Cite this