Search papers, labs, and topics across Lattice.
DuoGesture, a novel dual-stream approach, decomposes co-speech gesture generation into semantic and beat streams, coordinated by a Semantic Variational Information Bottleneck that selectively activates semantic gestures. Motion-Grounded Semantic Conditioning provides motion-aligned semantic priors, while an Inertial Beat Prior regularizes the beat stream, enhancing rhythmic consistency. Experiments demonstrate that DuoGesture surpasses holistic baselines in gesture quality, with ablations validating the contributions of each component.
Decomposing co-speech gesture generation into separate semantic and beat streams, controlled by a learned gating mechanism, yields more realistic and expressive human-computer interaction.
Co-speech gesture generation requires both semantic expressivity and biomechanically plausible rhythmic motion. Existing holistic gesture models mix lexically grounded semantic gestures with frequent prosody-aligned beat gestures. This limits semantic grounding, speech-motion alignment, and kinematic smoothness. We propose \emph{DuoGesture}, a neuro-inspired and biomechanically informed dual-stream approach that decomposes co-speech gesture synthesis into coupled semantic and beat streams. The two streams are coordinated by a \emph{Semantic Variational Information Bottleneck}, a stochastic frame-level gate that learns when semantic gestures should override rhythmic beat motion. The semantic stream is controlled by \emph{Motion-Grounded Semantic Conditioning}, which replaces purely linguistic word embeddings with motion-language representations to provide motion-aligned semantic priors for long-tailed lexical triggers of gestures. The beat stream is further regularised by an \emph{Inertial Beat Prior}, an anthropometry-weighted arm-chain module that reduces jitter and improves rhythmic consistency without constraining semantic frames. Objective evaluations and subjective experiments show that DuoGesture outperforms strong holistic baselines, while component ablations confirm the complementary roles of semantic grounding, stochastic stream selection, and biomechanical regularisation.
