The article discusses the challenges in collecting cross-age facial images, which limits the versatility of supervised methods for age-invariant face recognition, a critical task in applications such as security and biometrics. To address this issue, a novel semi-supervised learning approach named Cross-Age Contrastive Learning (CACon) is proposed. The method leverages an additional synthesized sample from the input image, introducing a new contrastive learning method. A new loss function in association with CACon is also proposed to perform contrastive learning on a triplet of samples. The paper shows that this method achieves state-of-the-art performance in homogeneous-dataset experiments on several age-invariant face recognition benchmarks and also outperforms other methods in cross-dataset experiments.
Publication date: 19 Dec 2023
Project Page: Not provided
Paper: https://arxiv.org/pdf/2312.11195
