Facial expression recognition depends on the detection of a few subtle facial feature traces. EMFACS (Emotion Facial Action Coding System) is a taxonomy of face muscle movements and positions called Action Units (AU) . AUs can be combined to describe complex facial expressions. We propose to (1) deconstruct facial expressions into face regions, grouping AUs by their proximity and contour direction; (2) recognize facial expressions by combining sparse reconstruction methods with face regions. We aim at finding a minimal set of AU to represent a given expression and apply l1 reconstruction to compute the deviation from the average face as an additive model of facial micro-expressions (the AUs). We compared our proposal to existing methods on the CK+  and JAFFE datasets . Our experiments indicate that sparse reconstruction with l1 penalty outperforms SVM and k-NN baselines. On the CK+ dataset, the best accuracy (89.8%) was obtained using sparse reconstruction.
|Title of host publication||Signal Processing Conference (EUSIPCO), 2014 Proceedings of the 22nd European|
|Publication status||Published - 10 Nov 2014|
|Event||EUSIPCO - |
Duration: 1 Jan 2014 → …
|Period||1/01/14 → …|