Collaborative similarity metric learning for face recognition in the wild

Abstract

Utilising different representations of face images is known to be helpful in face recognition. In this study, the authors propose two fusion techniques that make use of multiple face image features by collaboratively training a similarity metric learner, based on Siamese neural networks. This training procedure takes two (or possibly more) features of two face images and outputs a similarity score that depicts whether the faces belong to the same person or not. The authors investigate two approaches of collaborative similarity metric learning (CoSiM), both of which are based on training Siamese neural networks jointly, as a means of early fusion. The experiments are employed on hand-crafted features such as scale-invariant feature transform (SIFT) and variants of the local binary pattern (LBP), on the YouTube Faces and the Labeled Faces in the Wild data sets. The authors provide theoretical and empirical comparisons of the proposed models against the related methods in the literature. It is shown that the proposed technique improves on the verification accuracy, compared to single feature-based baselines. By only utilising simple features like SIFT and LBP, the proposed techniques are shown to yield comparable results to the state of the art techniques, which depend on deep convolutional architectures or higher level features.