Robust extraction of 3D facial nose feature
Yepeng Guan · Jisuanji gongcheng yu sheji · 2008
Aiming at the insufficiency of the current methods on detecting and locating nose based on large 3D facial data,a method to locate the nose automatically is developed.A 3D facial model is derived from mapping mesh of Delaunay triangulations to 3D space.The rough nose region is determined based on nose with concavo-convex characteristic.After filling holes and using the morphological filter algorithms to smooth the extracted nose region,the nose feature is determined.The experimental results are showed that the proposed algorithm is efficient and robust by extraction of nose feature across some real different scattered 3D facial data.