Year: 2024
Type: preprint
Abstract: We introduce a method for transferring the signer's appearance in sign language skeletal poses while preserving the sign content. Using estimated poses, we transfer the appearance of one signer to another, maintaining natural movements and transitions. This approach improves pose-based rendering and sign stitching while obfuscating identity. Our experiments show that while the method reduces signer identification accuracy, it slightly harms sign recognition performance, highlighting a tradeoff between privacy and utility. Our code is available at https://github.com/sign-language-processing/pose-anonymization.
Source: arXiv (Cornell University)
Authors: Moryossef, Amit, Sant, Gerard, Jiang, Zifan
Institution: University of Zurich
Language: English
Subfield: Human-Computer Interaction
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
Funder: Universität Zürich
OpenAlex ID: W4403579979