DiffSLT: Enhancing diversity in sign language translation via diffusion model

Download PDF
Open Webpage

JiHwan Moon, Jihoon Park, Jungeun Kim, Jongseong Bae, Hyeongwoo Jeon, Ha Young Kim

Published: 2025, Last Modified: 11 Nov 2025Pattern Recognit. Lett. 2025EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•Diffusion model in sign language translation offers diverse and accurate translations.•Multi-level visual features guide the diffusion process for better sign understanding.•Pseudo-gloss guidance further improves accuracy and bridges the cross-modal gap.•Extensive experiments show improved translation accuracy and greater lexical variety.