Markov Random Field Structures for Facial Action Unit Intensity Estimation
Abstract
We present a novel Markov Random Field (MRF) structure-based approach to the problem of facial action unit (AU) intensity estimation. AUs generally appear in common combinations, and exhibit strong relationships be-tween the intensities of a number of AUs. The aim of this work is to harness these links in order to improve the esti-mation of the intensity values over that possible from regres-sion of individual AUs. Our method exploits Support Vector Regression outputs to model appearance likelihoods of each individual AU, and integrates these with intensity combina-tion priors in MRF structures to improve the overall inten-sity estimates. We demonstrate the benefits of our approach on the upper face AUs annotated in the DISFA database, with significant improvements seen in both correlation and error rates for the majority of AUs, and on average. 1.
Cite
Text
Sandbach et al. "Markov Random Field Structures for Facial Action Unit Intensity Estimation." IEEE/CVF International Conference on Computer Vision Workshops, 2013. doi:10.1109/ICCVW.2013.101Markdown
[Sandbach et al. "Markov Random Field Structures for Facial Action Unit Intensity Estimation." IEEE/CVF International Conference on Computer Vision Workshops, 2013.](https://mlanthology.org/iccvw/2013/sandbach2013iccvw-markov/) doi:10.1109/ICCVW.2013.101BibTeX
@inproceedings{sandbach2013iccvw-markov,
title = {{Markov Random Field Structures for Facial Action Unit Intensity Estimation}},
author = {Sandbach, Georgia and Zafeiriou, Stefanos and Pantic, Maja},
booktitle = {IEEE/CVF International Conference on Computer Vision Workshops},
year = {2013},
pages = {738-745},
doi = {10.1109/ICCVW.2013.101},
url = {https://mlanthology.org/iccvw/2013/sandbach2013iccvw-markov/}
}