3D Face Recognition under Expressions, Occlusions, and Pose Variations

Abstract: We propose framework for analysing 3D faces, with the specific goals of comparing, matching, and averaging their shapes. Here we represent facial surfaces by radial curves emanating from the nose tips shape analysis. Here we detect facial deformations and large facial expressions large pose variations, missing parts, and partial occlusions due to glasses, hair, and illustrates the use of radial facial curves on 3D meshes. Here we represent facial surface by indexed collection of radial geodesic curves on 3D face meshes emanating from nose tip to the boundary of mesh and eyes and compare the facial shapes by comparing shapes of their corresponding curves. Here we using MATLAB for implementing our project.

Keywords: Occlusion- Blockage due to which image can not be properly seen. Pose Variation- Different views of the same image. Euclidean Distance- Distance between two points in Euclidean space. Face Recognition - a type of biometric software application that can identify a specific individual in a digital image by analysing and comparing patterns.