This repository contains demos for DiffSLVA：Harnessing Diffusion Models for Sign Language Video Anonymization
We propose text-guided zero-shot sign language video anonymization.
(1) Input prompt for anonymization: a superman in blue uniform is making gestures
(2) Input prompt for anonymization: a handsome man in CG style, blonde hair, is making gestures
(3) Input prompt for anonymization: a handsome man in CG style, blonde hair, is making gestures
The original signer is on the left is producing the ASL sentences with glosses, hand shapes, non-manual grammatical markings, and translations as shown below, available at https://dai.cs.rutgers.edu/dai/s/dai :
(1) and (2) - taken from the ASLLRP file Cory_2013-6-27_sc115, Utterance 22
(3) - taken from the ASLLRP file Cory_2013-6-27_sc113, Utterance 28
