Second Edition FRCSyn Challenge at CVPR 2024: Face Recognition Challenge in the Era of Synthetic Data

Synthetic data is gaining increasing relevance for training machine learningmodels. This is mainly motivated due to several factors such as the lack ofreal data and intra-class variability, time and errors produced in manuallabeling, and in some cases privacy concerns, among others. This paper presentsan overview of the 2nd edition of the Face Recognition Challenge in the Era ofSynthetic Data (FRCSyn) organized at CVPR 2024. FRCSyn aims to investigate theuse of synthetic data in face recognition to address current technologicallimitations, including data privacy concerns, demographic biases,generalization to novel scenarios, and performance constraints in challengingsituations such as aging, pose variations, and occlusions. Unlike the 1stedition, in which synthetic data from DCFace and GANDiffFace methods was onlyallowed to train face recognition systems, in this 2nd edition we propose newsub-tasks that allow participants to explore novel face generative methods. Theoutcomes of the 2nd FRCSyn Challenge, along with the proposed experimentalprotocol and benchmarking contribute significantly to the application ofsynthetic data to face recognition.