We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.GR

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Graphics

Title: ZeroEGGS: Zero-shot Example-based Gesture Generation from Speech

Abstract: We present ZeroEGGS, a neural network framework for speech-driven gesture generation with zero-shot style control by example. This means style can be controlled via only a short example motion clip, even for motion styles unseen during training. Our model uses a Variational framework to learn a style embedding, making it easy to modify style through latent space manipulation or blending and scaling of style embeddings. The probabilistic nature of our framework further enables the generation of a variety of outputs given the same input, addressing the stochastic nature of gesture motion. In a series of experiments, we first demonstrate the flexibility and generalizability of our model to new speakers and styles. In a user study, we then show that our model outperforms previous state-of-the-art techniques in naturalness of motion, appropriateness for speech, and style portrayal. Finally, we release a high-quality dataset of full-body gesture motion including fingers, with speech, spanning across 19 different styles.
Subjects: Graphics (cs.GR); Machine Learning (cs.LG); Sound (cs.SD)
Cite as: arXiv:2209.07556 [cs.GR]
  (or arXiv:2209.07556v2 [cs.GR] for this version)

Submission history

From: Saeed Ghorbani [view email]
[v1] Thu, 15 Sep 2022 18:34:30 GMT (30726kb,D)
[v2] Fri, 23 Sep 2022 20:49:15 GMT (30729kb,D)

Link back to: arXiv, form interface, contact.