GLOS: Sign Language Generation with Temporally Aligned Gloss-Level Conditioning

Taeryung Lee Hyeongjin Nam Gyeongsik Moon Kyoung Mu Lee · 2025

Read GLOS: Sign Language Generation with Temporally Aligned Gloss-Level Conditioning by Taeryung Lee Hyeongjin Nam Gyeongsik Moon Kyoung Mu Lee (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-07460.txt.

Description

Taeryung Lee, Hyeongjin Nam, Gyeongsik Moon, and Kyoung Mu Lee present a 2025 paper on making sign language from text. In the text itself, their system is called SIGNER. Sign language states meaning through a chain of gestures tied to word-like units called glosses. A good system must keep those gestures in the right order and make each one match its gloss.

Older methods blend all the text into one signal, which weakens the link between each gloss and its time slot. SIGNER first estimates the gloss chain and how long each gloss should last. It then spreads that meaning across time and blends it in locally, segment by segment. The goal is signed output with the right order and clearer meaning for each gesture.

Catalog record

CREATORTaeryung Lee Hyeongjin Nam Gyeongsik Moon Kyoung Mu Lee
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE33 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2506-07460

Browse the full library · About the Tower of Babel