On 3D face feature segmentation using implicit surface active contours
Matthias Krueger, Patrice Jean Delmas, Georgy Gimel’farb · 2008
Segmentation and extraction of face features from 2D images has been researched extensively in the past. Recent research shows that algorithms incorporating 3D information often perform better than purely 2D approaches in many areas of computer vision including face biometrics and recognition. Cheaper and readily available 3D data acquisition devices combined with improved 3D reconstruction algorithms delivering highly accurate data are leading a shift from 2D to 3D based data. In this paper we review our recently introduced algorithm for segmentation of 3D surfaces using active contours. Then we apply this framework to propose a novel method for feature segmentation from 3D faces. The performance of our model is illustrated on a set of 3D faces.