A Review on Partial Face Recognition System

Abstract

A number of face recognition methods are used for identification or verification of a person in many computer applications. Most of the methods use images of entire face under controlled condition for face recognition. However, in many situations the human faces might be occluded by other object which poses a difficulty in face recognition. To overcome this problem, the proposed approach identifies the persons. From partial face, the proposed system tries matching it with the full faces available in the database. It first extracts local features that are textural features and geometrical features and also detect keypoints. Then a Robust Point Set Matching (RPSM) method is used to match local features sets and also keypoints of the faces from database and the partial face image under investigation. A distance metric is used to find similarity of the two faces using their two sets of features.