A Critical Analysis on 3D Craniofacial Reconstruction Using GANs

3D Face and Head Reconstruction has gained a significant attention in various fields like biometrics, entertainment, medical imaging, and etc. Traditional reconstruction methods (which take information from number of 2D photos and geometrical shapes of the item) was less accurate in defining features of an object. Due to technological enhancement Generative Adversarial Networks (GANs) achieved a remarkable performance in finding the 3D face and head reconstruction features. This paper attempts to traverse the complexities of dataset selection and various hyperparameters of 3d craniofacial reconstruction in order to address the complex questions related to 3D face and head reconstruction using GANs and other approaches. This analysis draws on GANs as a general and proposed framework to provide insight into the fast-evolving field of 3D face and head reconstruction. To analyse the facts, 760 articles have been extracted based on Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) Guidelines. After a rigorous analysis, 38 articles have been finalized for final consideration. The findings of this study concludes that GANs can be used to recreate faces and heads in 3D using dense mesh structures that are highly textured and have distinct edges and vertices.