Abstract
Discrete motion tokenizers encode motion as atomic units and are widely used for co-speech gesture generation. It remains unclear which motion properties, especially those relevant to gesture semantics, are recoverable from these codebooks. We probe a reconstruction-trained codebook using 19 co-speech gesture descriptors spanning from raw motion to abstract communicative function. Results show that geometry and handedness are readily decodable from token embeddings, while motion category is only weakly decoded despite showing systematic differences in discrete code usage. This gap between reconstruction quality and descriptor decodability suggests that reconstruction objectives alone do not guarantee that gesture semantics are captured, and that evaluating codebooks on such properties can guide the design of more semantic motion tokenizers.
Keywords
Publication details
- Journal
- Not available
- Open access
- Green open access
Cite this article
APA 7
Suresh, V., Jain, D., Liu, J., Mughal, M. H., & Demberg, V. (2026). Do Motion Tokenizers for Co-Speech Gesture Generation Encode Gesture Semantics? https://omanscience.com/en/articles/do-motion-tokenizers-for-co-speech-gesture-generation-encode-gesture-semantics
MLA 9
Suresh, Varsha, et al. "Do Motion Tokenizers for Co-Speech Gesture Generation Encode Gesture Semantics?" https://omanscience.com/en/articles/do-motion-tokenizers-for-co-speech-gesture-generation-encode-gesture-semantics.
Chicago (author–date)
Suresh, Varsha, Divij Jain, Jia Liu, M. Hamza Mughal, and Vera Demberg. 2026. "Do Motion Tokenizers for Co-Speech Gesture Generation Encode Gesture Semantics?" https://omanscience.com/en/articles/do-motion-tokenizers-for-co-speech-gesture-generation-encode-gesture-semantics.
Harvard
Suresh, V., Jain, D., Liu, J., Mughal, M. H. and Demberg, V. (2026) 'Do Motion Tokenizers for Co-Speech Gesture Generation Encode Gesture Semantics?', Available at: https://omanscience.com/en/articles/do-motion-tokenizers-for-co-speech-gesture-generation-encode-gesture-semantics.
Vancouver
Suresh V, Jain D, Liu J, Mughal MH, Demberg V. Do Motion Tokenizers for Co-Speech Gesture Generation Encode Gesture Semantics? https://omanscience.com/en/articles/do-motion-tokenizers-for-co-speech-gesture-generation-encode-gesture-semantics
IEEE
V. Suresh, D. Jain, J. Liu, M. H. Mughal, and V. Demberg, "Do Motion Tokenizers for Co-Speech Gesture Generation Encode Gesture Semantics?," https://omanscience.com/en/articles/do-motion-tokenizers-for-co-speech-gesture-generation-encode-gesture-semantics.