Face Alignment Using a GAN-based Photorealistic Synthetic Dataset
Haoqi Gao, Koichi Ogawara · 2022
Face-related technology has matured over the past several decades. However, issues such as insufficient real-world training data and the privacy violations or data abuse caused by face applications have also triggered global controversy. For the question: "Can synthetic data be used to introduce novel variations in the real-world data? ". In this paper, we open a new research direction through synthetic datasets. We try to use synthetic datasets to reduce the dependence of the model on the real-world dataset. However, considering the differences between synthetic and real-world data, our work aims to convert the synthetic face images generated by the Face generating middleware 3D model (FaceGen) into more realistic face images for training face alignment algorithms.