Handling Data Imbalance in Automatic Facial Action Intensity Estimation

Philipp Werner, Frerk Saxen and Ayoub Al-Hamadi

Abstract

Automatic Action Unit (AU) intensity estimation is a key problem in facial expression analysis. But limited research attention has been paid to the inherent class imbalance, which usually leads to suboptimal performance. To handle the imbalance, we propose (1) a novel multiclass under-sampling method and (2) its use in an ensemble. We compare our approach with state of the art sampling methods used for AU intensity estimation. Multiple datasets and widely varying performance measures are used in the literature, making direct comparison difficult. To address these shortcomings, we compare different performance measures for AU intensity estimation and evaluate our proposed approach on three publicly available datasets, with a comparison to state of the art methods along with a cross dataset evaluation.

Session

Poster 2

Files

PDF iconExtended Abstract (PDF, 603K)
PDF iconPaper (PDF, 801K)
ZIP iconSupplemental Materials (ZIP, 2M)

DOI

10.5244/C.29.124
https://dx.doi.org/10.5244/C.29.124

Citation

Philipp Werner, Frerk Saxen and Ayoub Al-Hamadi. Handling Data Imbalance in Automatic Facial Action Intensity Estimation. In Xianghua Xie, Mark W. Jones, and Gary K. L. Tam, editors, Proceedings of the British Machine Vision Conference (BMVC), pages 124.1-124.12. BMVA Press, September 2015.

Bibtex

@inproceedings{BMVC2015_124,
	title={Handling Data Imbalance in Automatic Facial Action Intensity Estimation},
	author={Philipp Werner and Frerk Saxen and Ayoub Al-Hamadi},
	year={2015},
	month={September},
	pages={124.1-124.12},
	articleno={124},
	numpages={12},
	booktitle={Proceedings of the British Machine Vision Conference (BMVC)},
	publisher={BMVA Press},
	editor={Xianghua Xie, Mark W. Jones, and Gary K. L. Tam},
	doi={10.5244/C.29.124},
	isbn={1-901725-53-7},
	url={https://dx.doi.org/10.5244/C.29.124}
}