Surface shape descriptors on 3D faces

Abstract EN

The general objective of surface shape descriptors techniques is to categorize several surface shapes from collection data.

Gaussian (K) and Mean (H) curvatures are the most broadly utilized indicators for surface shape characterization in collection image analysis.

This paper explains the details of some descriptions (K and H), The discriminating power of 3D descriptors taken away from 3D surfaces (faces) is analyzed and present the experiment results of applying these descriptions on 3D face (with polygon mesh and point cloud representations).

The results shows that Gaussian and Mean curvatures are important to discover unique points on the 3d surface (face) and the experiment result shows that these curvatures are very useful for some special features extraction for eyes and nose detection.