Integrated Speech and Gesture Synthesis (ICMI 2021)

paper link: https://arxiv.org/pdf/2108.11436.pdf
code: https://github.com/swatsw/isg_official

videos:

See paper sec.3 & sec.5 for details of each model. "ST-Tacotron2-ISG-using_mel_output" is an ablation model discussed in sec.7.
All input texts are generated by a fine-tuned GPT-2 for the sole purpose of evaluating the models (details in paper sec. 4.2).



***
All rights reserved by authors of the paper "Integrated Speech and Gesture Synthesis"(ICMI 2021).
All videos are created by the authors.
These are for personal viewing and for academic research purpose only.
Redistribution or reuse of any part of any material shown on this website or in the paper is prohibited.