pi-GAN: Periodic Implicit Generative Adversarial Networks for 3D-Aware Image Synthesis
This work provides a significant improvement in the quality and 3D consistency of synthesized images for researchers and practitioners working on generative models and neural rendering.
This paper addresses the limitations of existing 3D-aware image synthesis methods, which often suffer from view inconsistency and insufficient image quality due to unexpressive network architectures. The authors propose pi-GAN, a novel generative model that uses neural representations with periodic activation functions and volumetric rendering to create view-consistent 3D representations with fine detail, achieving state-of-the-art results on multiple datasets.
We have witnessed rapid progress on 3D-aware image synthesis, leveraging recent advances in generative visual models and neural rendering. Existing approaches however fall short in two ways: first, they may lack an underlying 3D representation or rely on view-inconsistent rendering, hence synthesizing images that are not multi-view consistent; second, they often depend upon representation network architectures that are not expressive enough, and their results thus lack in image quality. We propose a novel generative model, named Periodic Implicit Generative Adversarial Networks ($π$-GAN or pi-GAN), for high-quality 3D-aware image synthesis. $π$-GAN leverages neural representations with periodic activation functions and volumetric rendering to represent scenes as view-consistent 3D representations with fine detail. The proposed approach obtains state-of-the-art results for 3D-aware image synthesis with multiple real and synthetic datasets.