Age-Invariant Face Embedding using the Wasserstein Distance

Eran Dahan,Yosi Keller

DOI: https://doi.org/10.48550/arXiv.2305.02745

2023-05-04

Abstract:In this work, we study face verification in datasets where images of the same individuals exhibit significant age differences. This poses a major challenge for current face recognition and verification techniques. To address this issue, we propose a novel approach that utilizes multitask learning and a Wasserstein distance discriminator to disentangle age and identity embeddings of facial images. Our approach employs multitask learning with a Wasserstein distance discriminator that minimizes the mutual information between the age and identity embeddings by minimizing the Jensen-Shannon divergence. This improves the encoding of age and identity information in face images and enhances the performance of face verification in age-variant datasets. We evaluate the effectiveness of our approach using multiple age-variant face datasets and demonstrate its superiority over state-of-the-art methods in terms of face verification accuracy.

Computer Vision and Pattern Recognition