Disentangled Dynamic Representations from Unordered Data
Machine Learning
2018-12-11 v1 Machine Learning
Abstract
We present a deep generative model that learns disentangled static and dynamic representations of data from unordered input. Our approach exploits regularities in sequential data that exist regardless of the order in which the data is viewed. The result of our factorized graphical model is a well-organized and coherent latent space for data dynamics. We demonstrate our method on several synthetic dynamic datasets and real video data featuring various facial expressions and head poses.
Cite
@article{arxiv.1812.03962,
title = {Disentangled Dynamic Representations from Unordered Data},
author = {Leonhard Helminger and Abdelaziz Djelouah and Markus Gross and Romann M. Weber},
journal= {arXiv preprint arXiv:1812.03962},
year = {2018}
}
Comments
Symposium on Advances in Approximate Bayesian Inference, 2018