In this paper we investigate the combination of shape features and Phase-based Gabor features for context-independent Action Unit Recognition. For our recognition goal, three regions of interest have been devised that efficiently capture the AUs activation/deactivation areas. In each of these regions a feature set consisting of geometrical and histogram of Gabor phase appearance-based features have been estimated. For each Action Unit, we applied Adaboost for feature selection, and used a binary SVM for context-independent classification. Using the Cohn-Kanade database, we achieved an average F 1 score of 93.8% and an average area under the ROC curve of 97.9 %, for the 11 AUs considered.
Gonzalez, I, Sahli, H, Enescu, V & Verhelst, W 2011, Context-Independent Facial Action Unit Recognition Using Shape and Gabor Phase Information. in S D?mello, A Graesser, B Schuller & J Martin (eds), Affective Computing and Intelligent Interaction. vol. 6974, Lecture Notes in Computer Science, Springer, pp. 548-557, Unknown, 1/01/11. <http://www.springerlink.com/content/2265227761271337/>
Gonzalez, I., Sahli, H., Enescu, V., & Verhelst, W. (2011). Context-Independent Facial Action Unit Recognition Using Shape and Gabor Phase Information. In S. D?mello, A. Graesser, B. Schuller, & J. Martin (Eds.), Affective Computing and Intelligent Interaction (Vol. 6974, pp. 548-557). (Lecture Notes in Computer Science). Springer. http://www.springerlink.com/content/2265227761271337/
@inproceedings{de8cea8b0ec5447a89edb9b67e53a46d,
title = "Context-Independent Facial Action Unit Recognition Using Shape and Gabor Phase Information",
abstract = "In this paper we investigate the combination of shape features and Phase-based Gabor features for context-independent Action Unit Recognition. For our recognition goal, three regions of interest have been devised that efficiently capture the AUs activation/deactivation areas. In each of these regions a feature set consisting of geometrical and histogram of Gabor phase appearance-based features have been estimated. For each Action Unit, we applied Adaboost for feature selection, and used a binary SVM for context-independent classification. Using the Cohn-Kanade database, we achieved an average F 1 score of 93.8% and an average area under the ROC curve of 97.9 %, for the 11 AUs considered.",
keywords = "Computer Science",
author = "Isabel Gonzalez and Hichem Sahli and Valentin Enescu and Werner Verhelst",
note = "Sidney D?Mello, Arthur Graesser, Bj{\"o}rn Schuller, Jean-Claud Martin; Unknown ; Conference date: 01-01-2011",
year = "2011",
language = "English",
isbn = "978-3-642-24599-2",
volume = "6974",
series = "Lecture Notes in Computer Science",
publisher = "Springer",
pages = "548--557",
editor = "Sidney D?mello and Arthur Graesser and Bj{\"o}rn Schuller and Jean-claud Martin",
booktitle = "Affective Computing and Intelligent Interaction",
}