An effective method of 3D facial features segmentation

The facial features segmentation based on image information, due to the influence of illumination, posture variation and expression, segmentation accuracy is literally not too inspiring. Using 3D face model is one of the effective approaches to tackle the problems above. In this paper, an effective method of 3D facial features segmentation based on the geometric features and statistical features of the 3D coordinates of the face model is proposed. The key curvatures, such as gaussian and mean curvature vector, maximum and minimum curvature vector, are first extracted based on triangular mesh surface analysis. Then, 3D facial feature regions are roughly partitioned by the method of clustering. The optimizing segmentation result is subsequently completed under the guidance of the structural model of facial feature region which is built based on diverse structural characteristics of a human face. Experimental results demonstrate that our method has an excellent performance of the segmentation of the eyes, nose, and mouth regions.