Young Labeled Faces in the Wild (YLFW): A Dataset for Children Faces Recognition

TL;DR

This paper introduces YLFW, a new standardized dataset for children's face recognition, addressing biases in existing datasets and providing a benchmark for evaluating models on diverse, age-specific face recognition challenges.

Contribution

The paper presents the first standardized dataset and development set for children's face recognition, covering diverse races and enabling benchmarking and model adaptation.

Findings

Benchmark dataset for children's face recognition introduced.

Experiments demonstrate the dataset's effectiveness for model evaluation.

Addresses bias towards adult faces in existing datasets.

Abstract

Face recognition has achieved outstanding performance in the last decade with the development of deep learning techniques. Nowadays, the challenges in face recognition are related to specific scenarios, for instance, the performance under diverse image quality, the robustness for aging and edge cases of person age (children and elders), distinguishing of related identities. In this set of problems, recognizing children's faces is one of the most sensitive and important. One of the reasons for this problem is the existing bias towards adults in existing face datasets. In this work, we present a benchmark dataset for children's face recognition, which is compiled similarly to the famous face recognition benchmarks LFW, CALFW, CPLFW, XQLFW and AgeDB. We also present a development dataset (separated into train and test parts) for adapting face recognition models for face images of children. The proposed data is balanced for African, Asian, Caucasian, and Indian races. To the best of our knowledge, this is the first standartized data tool set for benchmarking and the largest collection for development for children's face recognition. Several face recognition experiments are presented to demonstrate the performance of the proposed data tool set.