With the ever-growing demand for large datasets in deep learning, the problem of data scarcity arises during the implementation of these models. This leads to the Generative Adversarial Networks (GAN) as a promising solution among researchers, which has the capability of producing new data using the same statistics as the training set. In this paper, GAN networks are analyzed for image-based datasets and for producing realistic images that appear genuine to human observers. Furthermore, the Deep Convolutional Generative Adversarial Network (DCGAN) is explored for artificial face image generation. DCGANs represent a powerful class of generative models capable of producing high-fidelity images. Thus, the results of DCGAN on celeb AI datasets are presented in this paper to show the efficacy of the model for synthetic face image generation. The work presented here demonstrates the ability of DCGANs to generate realistic and visually compelling images, often indistinguishable from their real-world counterparts. Its performance surpasses many alternative models trained on similar datasets, highlighting the efficacy of DCGANs in this domain. Furthermore, the success of this study opens avenues for exploring the potential of DCGANs in various applications in diverse fields such as computer vision, computer graphics, artistic creation, etc. This research, with its extensive exploration of DCGANs as a solution to data scarcity, offers valuable insights with significant implications for future advancements in these areas.