Utilization of Generative Adversarial Networks in Face Image Synthesis for Augmentation of Face Recognition Training Data

Face recognition has become a popular research field in computer vision and is widely applied in various sectors. The challenge with face recognition is that if the training data is limited, the face recognition rate will be less effective. Generative Adversarial Networks (GANs) is a deep learning method that can create synthesis images with high quality. This research aims to utilize GANs in synthesizing face images as a form of augmentation in face recognition training data. Initially, the latent space representation of the face image will be made using GANs, then adding styles to the face image using the latent direction method. In the experiment of making latent space representation, the loss value was able to reach 0.15. In the experiment of face recognition, the addition of face image synthesis was able to increase the accuracy of the face recognition classifier model from 0.74 to 0.89.