Pose Tolerant Face Recognition: A Review

Abstract

One of the main hazards found in the face recognition techniques lies in the difficulties of manipulating varying poses. The facial image variations caused due to the rotations are often more notable than the inter-person variations, used in differentiating identities. Pose-invariant face recognition (PIFR) has high potentials in several applications that deal with the uncooperative subjects, in which the full potential of facial recognition being a quiet biometric technique that can be implemented and utilized. Extensive attempts have been set into the research towards PIFR over recent years, and many eminent strategies have been introduced. However, there are some issues in face recognition across pose that remain unsolved, such as lack of knowledge about the subspaces of pose variant images. This paper comes up with a crucial survey of researches on image-based face recognition across varying poses. The current techniques are holistically reviewed and presented.