PD-GS: Phoneme-Driven 3DGS for Audio-Driven Talking Heads

[Submitted on 5 Aug 2026]

View PDF HTML (experimental)

Abstract:3D Gaussian Splatting (3DGS) enables fast, photorealistic talking-head rendering, yet accurate lip articulation remains elusive: mouth motion is often over-smoothed and may violate hard articulatory constraints such as bilabial closures, producing the notorious “leaky mouth” artifact. A key difficulty is that brief, discrete articulatory events are inferred from a continuous acoustic embedding under a regression objective, which biases predictions toward averaged mouth configurations. While modern self-supervised speech encoders provide rich prosodic and phonetic cues, they do not provide an explicit, frame-aligned linguistic target that reliably disambiguates closure-level events. We propose \textbf{Phoneme-Driven Gaussian Splatting (PD-GS)}, which augments a 3DGS talker with time-aligned phoneme tokens obtained from an automatic ASR and forced-alignment pipeline. Our core component, the \textbf{Linguistic Fusion Module (LFM)}, adaptively fuses continuous audio context with discrete phoneme embeddings through a learned gate, allowing the model to preserve smooth audio-driven dynamics while strengthening phoneme guidance on articulation-critical segments. PD-GS is trained purely from monocular video using image reconstruction and lip landmark supervision. On HDTF, PD-GS achieves the best lip geometry among the compared baselines (LMD 2.66) and qualitatively reduces closure violations in challenging phoneme sequences, yielding more linguistically faithful neural avatars.

Submission history

From: Ao Fu [view email]
[v1] Wed, 5 Aug 2026 10:31:42 UTC (4,961 KB)