Back to Search Start Over

Speech-driven facial animation using polynomial fusion of features

Authors :
Kefalas, Triantafyllos
Vougioukas, Konstantinos
Panagakis, Yannis
Petridis, Stavros
Kossaifi, Jean
Pantic, Maja
Publication Year :
2019

Abstract

Speech-driven facial animation involves using a speech signal to generate realistic videos of talking faces. Recent deep learning approaches to facial synthesis rely on extracting low-dimensional representations and concatenating them, followed by a decoding step of the concatenated vector. This accounts for only first-order interactions of the features and ignores higher-order interactions. In this paper we propose a polynomial fusion layer that models the joint representation of the encodings by a higher-order polynomial, with the parameters modelled by a tensor decomposition. We demonstrate the suitability of this approach through experiments on generated videos evaluated on a range of metrics on video quality, audiovisual synchronisation and generation of blinks.

Details

Database :
arXiv
Publication Type :
Report
Accession number :
edsarx.1912.05833
Document Type :
Working Paper