Kinship verification from faces: Methods, databases and challenges

Abstract

Automatic kinship verification from face images aims at verifying whether two persons have a kin relation or not by only comparing their visual facial similarities. This is an emerging topic in face analysis and computer vision. Automatic kinship verification is useful in many applications such as finding missing children, image annotation, and social media analysis. In the recent years, there has been a lot of efforts in developing methods to cope with this challenging but important research problem. This paper introduces the topic, discusses the proposed methods, describes the available databases, highlights the open issues and challenges and points out directions for future research.