Kinship verification using multiview hybrid distance learning

The estimation of kin relationships between parents and their children based on their face images is a common biometric task, conducted daily by human observers. Kin similarity is subject to significant appearance variability, as parents and their children differ by age and gender. In this work we propose a multiview hybrid combined symmetric and asymmetric distance learning network for facial kinship verification. Dual discriminative representations are learnt for the parents and the children using a margin maximization learning scheme, while the kin verification is formulated as a classification problem solved by SVM. The proposed scheme was successfully applied to the KinFaceW and KinFaceCornell datasets, comparing favorably with contemporary state-of-the-art approaches.