Talk3D: High-Fidelity Talking Portrait Synthesis via Personalized 3D Generative Prior

TL;DR

Talk3D introduces a novel framework for high-fidelity, audio-driven talking head synthesis that accurately reconstructs facial geometry using a personalized 3D generative prior, outperforming existing methods especially in extreme poses.

Contribution

The paper proposes a new audio-guided attention U-Net architecture that leverages a pre-trained 3D generative prior to improve facial geometry reconstruction in talking head synthesis.

Findings

Outperforms state-of-the-art benchmarks in realism and accuracy.

Excels in generating facial geometries under extreme head poses.

Effectively disentangles audio-related and unrelated facial variations.

Abstract

Recent methods for audio-driven talking head synthesis often optimize neural radiance fields (NeRF) on a monocular talking portrait video, leveraging its capability to render high-fidelity and 3D-consistent novel-view frames. However, they often struggle to reconstruct complete face geometry due to the absence of comprehensive 3D information in the input monocular videos. In this paper, we introduce a novel audio-driven talking head synthesis framework, called Talk3D, that can faithfully reconstruct its plausible facial geometries by effectively adopting the pre-trained 3D-aware generative prior. Given the personalized 3D generative model, we present a novel audio-guided attention U-Net architecture that predicts the dynamic face variations in the NeRF space driven by audio. Furthermore, our model is further modulated by audio-unrelated conditioning tokens which effectively disentangle variations unrelated to audio features. Compared to existing methods, our method excels in generating realistic facial geometries even under extreme head poses. We also conduct extensive experiments showing our approach surpasses state-of-the-art benchmarks in terms of both quantitative and qualitative evaluations.