EMOPortraits: Emotion-enhanced Multimodal One-shot Head Avatars

TL;DR

This paper introduces EMOPortraits, an advanced multimodal head avatar model that enhances expression fidelity, supports speech-driven animation, and is evaluated on a new dataset with diverse facial expressions.

Contribution

The paper proposes a new model, EMOPortraits, with architectural and training improvements to better capture intense expressions and integrate speech-driven animation, outperforming previous models.

Findings

Sets a new state-of-the-art in emotion transfer accuracy.

Achieves top performance in audio-driven facial animation.

Introduces a novel multi-view video dataset with diverse expressions.

Abstract

Head avatars animated by visual signals have gained popularity, particularly in cross-driving synthesis where the driver differs from the animated character, a challenging but highly practical approach. The recently presented MegaPortraits model has demonstrated state-of-the-art results in this domain. We conduct a deep examination and evaluation of this model, with a particular focus on its latent space for facial expression descriptors, and uncover several limitations with its ability to express intense face motions. To address these limitations, we propose substantial changes in both training pipeline and model architecture, to introduce our EMOPortraits model, where we: Enhance the model's capability to faithfully support intense, asymmetric face expressions, setting a new state-of-the-art result in the emotion transfer task, surpassing previous methods in both metrics and quality. Incorporate speech-driven mode to our model, achieving top-tier performance in audio-driven facial animation, making it possible to drive source identity through diverse modalities, including visual signal, audio, or a blend of both. We propose a novel multi-view video dataset featuring a wide range of intense and asymmetric facial expressions, filling the gap with absence of such data in existing datasets.