Supervised Contrastive Learning and Feature Fusion for Improved Kinship Verification

TL;DR

This paper introduces a supervised contrastive learning approach combined with feature fusion to enhance facial kinship verification, achieving state-of-the-art accuracy on the FIW dataset.

Contribution

It proposes a novel supervised contrastive learning method specifically designed for kinship verification, improving accuracy over previous approaches.

Findings

Achieved 81.1% accuracy on FIW dataset.

Outperformed existing kinship verification methods.

Demonstrated the effectiveness of feature fusion with contrastive learning.

Abstract

Facial Kinship Verification is the task of determining the degree of familial relationship between two facial images. It has recently gained a lot of interest in various applications spanning forensic science, social media, and demographic studies. In the past decade, deep learning-based approaches have emerged as a promising solution to this problem, achieving state-of-the-art performance. In this paper, we propose a novel method for solving kinship verification by using supervised contrastive learning, which trains the model to maximize the similarity between related individuals and minimize it between unrelated individuals. Our experiments show state-of-the-art results and achieve 81.1% accuracy in the Families in the Wild (FIW) dataset.