Publication details

Facial expression recognition using angle-related information from facial meshes

Conference Paper (international conference)

Vretos N., Solachidis V., Somol Petr, Pitas I.

serial: Proc. 16th European Signal Processing Conference (EUSIPCO- 2008), p. 1-5

action: 16th European Signal Processing Conference (EUSIPCO- 2008), (Lausanne, CH, 25.08.2008-29.08.2008)

research: CEZ:AV0Z10750506

project(s): 1M0572, GA MŠk, GA102/08/0593, GA ČR, 507752,

keywords: facial expression, facial meshes, recognition, feature selection

preview: Download

abstract (eng):

In this paper we introduce a new method for facial expression recognition. In order to be able to recognize the six main facial expressions we use a grid approach and therefore we establish our new feature space based on the angles that each grid’s edge form. This way we undertake several affine transformations such as translation, rotation and scaling which in other approaches are considered very harmful in the overall accuracy of a facial expression recognition algorithm. We will therefore demonstrate how we create this feature space, as well as how we apply a feature selection process within this space. The angular nature of the data impose some considerations which will be clarified in this paper.

abstract (cze):

Prezentujeme novou metodu rozpoznávání výrazu tváře. Abychom mohli rozpoznat jeden z šesti základních typů výrazů, aplikujeme síť záchytných bodů a tím vytváříme nový prostor popisu problému. Tím podchycujeme afinní transformace které u jiných přístupů jsou zdrojem problémů. V novém prostoru provádíme výběr příznaků.