Novel approach to Create Human Faces with DCGAN for Face Recognition

Abstract

Due to the remarkable data generation abilities of the generative models, many generative adversarial networks (GAN) models have been developed, and several real-world applications in computer vision and machine learning have emerged. The generative models have received significant attention in the field of unsupervised learning via this new and useful framework. In spite of GAN's outstanding performance, steady training remains a challenge. In this model, use of Deep Convolutional Generative Adversarial Networks is incorporated, Main aim is to produce human faces from unlabeled data. Face generation has a wide range of applications in image processing, entertainment, and other industries. Extensive simulation is performed on the CelebA     dataset. Key focus is to successfully construct human faces from the unlabeled data and random noise and achieved average loss of 1.115% and 0.5894 % for generator and discriminator respectively.