Radial curve-based classification and localization of features for 3D facial expressions
Manar D. Samad, Khan M. Iftekharuddin · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2014
We present a novel method to analyze facial expressions from 3D shape using radial curves and curve based geometric features. Curve based representation of 3D facial shape and corresponding geometric features overcome the curse of dimensionality providing a means for fast and automatic classification and comparison of 3D facial shapes. Our proposed curve based geometric features effectively capture local variations and classify facial expressions from 3D facial shapes. A multiclass feature selection technique is used to identify the most effective features that localize the effective regions of the face. Six basic facial expressions are classified using a publicly available 3D facial expression dataset.