Residual diffusion improves multiple facial reaction generation accuracy
ResDiffFRG: Residual Diffusion for Multiple Appropriate Facial Reaction Generation
Machine Learning
Summary
In conversations, listeners show facial reactions that help speakers understand their feelings. Because people often have many natural ways to react, it's important for computers to generate multiple suitable facial reactions for virtual listeners. The authors propose ResDiffFRG, a method that bases its facial reaction generation on the speaker's face, so it only needs to create the small changes that make a proper listener reaction. This approach improves how well generated reactions match real human responses compared to earlier methods.
What this means in practice
- •For virtual agent developers: Generate multiple realistic listener facial reactions aligned with speaker behaviour to create more natural conversational agents.$Commercial implications: Enables virtual assistants and customer service bots to display diverse and context-appropriate facial reactions, enhancing user engagement and satisfaction.
- •For game animation teams: Produce varied and appropriate facial responses for non-player characters during dialogues to improve realism and player immersion.
Authors
Shizhe Liu, Jiayan Gu, Xiangyu Kong, Siyang Song
Abstract
In dyadic human speaker-listener conversations, the listener's facial reactions allows the speaker to accurately perceive the listener's emotional states. Since human facial reactions are non-deterministic, the ability to generate multiple appropriate human-like facial reactions is crucial for realistic human-agent interactions. Although diffusion models are naturally suited to such one-to-many generation, existing diffusion-based Multiple Appropriate Facial Reaction Generation (MAFRG) methods attempt to denoise random Gaussian initialisations directly into multiple appropriate facial reactions (AFRs). These random initialisations are usually not well-aligned with the target listener facial reaction, which requires complex denoising trajectories from these initialisations, and subsequently creates substantial opportunities for deviations away from the range of trajectories leading to appropriate AFRs. Given the inherent mimicry between the human listener's and speaker's facial behaviours, we address the above denoising trajectory issue by leveraging this strong prior. Specifically, we propose ResDiffFRG, a novel diffusion-based MAFRG framework that explicitly anchors the diffusion process to the speaker behaviour by defining its diffusion target as the residual between the speaker anchor and an AFR. The denoiser only needs to model the comparatively small, reaction-specific residual needed to transform this anchor into an AFR, rather than reconstructing the complete reaction from an unstructured state. Extensive experiments show that ResDiffFRG achieves large improvements in correlation-based appropriateness over existing methods. Our denoising trajectory analysis showed that even at the start of the denoising trajectory, ResDiffFRG already achieves a higher facial-reaction correlation score than the Gaussian Diffusion baseline does after completing 60% of its denoising trajectory.