GenCA: A Text-conditioned Generative Model for Realistic and Drivable Codec Avatars

TL;DR

This paper introduces GenCA, a text-conditioned generative model that creates realistic, detailed, and controllable 3D avatars, overcoming limitations of traditional methods and enabling diverse applications like editing and reconstruction.

Contribution

We develop a novel text-conditioned generative model that produces high-fidelity, editable 3D avatars with complete facial details and robust drivability, advancing beyond existing static or limited-avatar approaches.

Findings

Generates diverse, photo-realistic 3D avatars from text prompts.

Enables controllable avatar expression and identity manipulation.

Supports downstream tasks like avatar editing and single-shot reconstruction.

Abstract

Photo-realistic and controllable 3D avatars are crucial for various applications such as virtual and mixed reality (VR/MR), telepresence, gaming, and film production. Traditional methods for avatar creation often involve time-consuming scanning and reconstruction processes for each avatar, which limits their scalability. Furthermore, these methods do not offer the flexibility to sample new identities or modify existing ones. On the other hand, by learning a strong prior from data, generative models provide a promising alternative to traditional reconstruction methods, easing the time constraints for both data capture and processing. Additionally, generative methods enable downstream applications beyond reconstruction, such as editing and stylization. Nonetheless, the research on generative 3D avatars is still in its infancy, and therefore current methods still have limitations such as creating static avatars, lacking photo-realism, having incomplete facial details, or having limited drivability. To address this, we propose a text-conditioned generative model that can generate photo-realistic facial avatars of diverse identities, with more complete details like hair, eyes and mouth interior, and which can be driven through a powerful non-parametric latent expression space. Specifically, we integrate the generative and editing capabilities of latent diffusion models with a strong prior model for avatar expression driving. Our model can generate and control high-fidelity avatars, even those out-of-distribution. We also highlight its potential for downstream applications, including avatar editing and single-shot avatar reconstruction.