CGOF++: Controllable 3D Face Synthesis with Conditional Generative Occupancy Fields

TL;DR

This paper introduces CGOF++, a novel 3D face synthesis framework that leverages conditional generative occupancy fields and 3D priors to produce high-fidelity, controllable face images with precise 3D shape control, surpassing 2D methods.

Contribution

The paper proposes a new NeRF-based 3D face synthesis method using cGOF++, integrating 3D priors and novel loss functions for improved 3D controllability and image quality.

Findings

Effective 3D controllability over face shapes.

High-fidelity face image generation.

Outperforms state-of-the-art 2D methods.

Abstract

Capitalizing on the recent advances in image generation models, existing controllable face image synthesis methods are able to generate high-fidelity images with some levels of controllability, e.g., controlling the shapes, expressions, textures, and poses of the generated face images. However, previous methods focus on controllable 2D image generative models, which are prone to producing inconsistent face images under large expression and pose changes. In this paper, we propose a new NeRF-based conditional 3D face synthesis framework, which enables 3D controllability over the generated face images by imposing explicit 3D conditions from 3D face priors. At its core is a conditional Generative Occupancy Field (cGOF++) that effectively enforces the shape of the generated face to conform to a given 3D Morphable Model (3DMM) mesh, built on top of EG3D [1], a recent tri-plane-based generative model. To achieve accurate control over fine-grained 3D face shapes of the synthesized images, we additionally incorporate a 3D landmark loss as well as a volume warping loss into our synthesis framework. Experiments validate the effectiveness of the proposed method, which is able to generate high-fidelity face images and shows more precise 3D controllability than state-of-the-art 2D-based controllable face synthesis methods.