Talk3D: High-Fidelity Talking Portrait Synthesis via Personalized 3D Generative Prior
This work addresses the challenge of generating high-fidelity, 3D-consistent talking portraits for applications like virtual avatars and video synthesis, representing a strong specific gain rather than a broad paradigm shift.
The paper tackles the problem of incomplete face geometry in audio-driven talking head synthesis from monocular videos by introducing Talk3D, which uses a personalized 3D generative prior and an audio-guided attention U-Net to predict dynamic face variations in NeRF space, resulting in realistic facial geometries even under extreme head poses and surpassing state-of-the-art benchmarks in evaluations.
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.