Segment, Embed, and Align: A Universal Recipe for Aligning Subtitles to Signing

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Jiang, Zifan, Jang, Youngjoon, Momeni, Liliane, Varol, Gül, Ebling, Sarah, Zisserman, Andrew
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866909950511415296
author Jiang, Zifan
Jang, Youngjoon
Momeni, Liliane
Varol, Gül
Ebling, Sarah
Zisserman, Andrew
author_facet Jiang, Zifan
Jang, Youngjoon
Momeni, Liliane
Varol, Gül
Ebling, Sarah
Zisserman, Andrew
contents The goal of this work is to develop a universal approach for aligning subtitles (i.e., spoken language text with corresponding timestamps) to continuous sign language videos. Prior approaches typically rely on end-to-end training tied to a specific language or dataset, which limits their generality. In contrast, our method Segment, Embed, and Align (SEA) provides a single framework that works across multiple languages and domains. SEA leverages two pretrained models: the first to segment a video frame sequence into individual signs and the second to embed the video clip of each sign into a shared latent space with text. Alignment is subsequently performed with a lightweight dynamic programming procedure that runs efficiently on CPUs within a minute, even for hour-long episodes. SEA is flexible and can adapt to a wide range of scenarios, utilizing resources from small lexicons to large continuous corpora. Experiments on four sign language datasets demonstrate state-of-the-art alignment performance, highlighting the potential of SEA to generate high-quality parallel data for advancing sign language processing. SEA's code and models are openly available.
format Preprint
id arxiv_https___arxiv_org_abs_2512_08094
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Segment, Embed, and Align: A Universal Recipe for Aligning Subtitles to Signing
Jiang, Zifan
Jang, Youngjoon
Momeni, Liliane
Varol, Gül
Ebling, Sarah
Zisserman, Andrew
Computation and Language
The goal of this work is to develop a universal approach for aligning subtitles (i.e., spoken language text with corresponding timestamps) to continuous sign language videos. Prior approaches typically rely on end-to-end training tied to a specific language or dataset, which limits their generality. In contrast, our method Segment, Embed, and Align (SEA) provides a single framework that works across multiple languages and domains. SEA leverages two pretrained models: the first to segment a video frame sequence into individual signs and the second to embed the video clip of each sign into a shared latent space with text. Alignment is subsequently performed with a lightweight dynamic programming procedure that runs efficiently on CPUs within a minute, even for hour-long episodes. SEA is flexible and can adapt to a wide range of scenarios, utilizing resources from small lexicons to large continuous corpora. Experiments on four sign language datasets demonstrate state-of-the-art alignment performance, highlighting the potential of SEA to generate high-quality parallel data for advancing sign language processing. SEA's code and models are openly available.
title Segment, Embed, and Align: A Universal Recipe for Aligning Subtitles to Signing
topic Computation and Language
url https://arxiv.org/abs/2512.08094