DiffSLT: Enhancing Diversity in Sign Language Translation via Diffusion Model

Sign language translation (SLT) is challenging, as it involves converting sign language videos into natural language. Previous studies have prioritized accuracy over diversity. However, diversity is crucial for handling lexical and syntactic ambiguities in machine translation, suggesting it could si...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Moon, JiHwan, Park, Jihoon, Kim, Jungeun, Bae, Jongseong, Jeon, Hyeongwoo, Kim, Ha Young
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Moon, JiHwan
Park, Jihoon
Kim, Jungeun
Bae, Jongseong
Jeon, Hyeongwoo
Kim, Ha Young
description Sign language translation (SLT) is challenging, as it involves converting sign language videos into natural language. Previous studies have prioritized accuracy over diversity. However, diversity is crucial for handling lexical and syntactic ambiguities in machine translation, suggesting it could similarly benefit SLT. In this work, we propose DiffSLT, a novel gloss-free SLT framework that leverages a diffusion model, enabling diverse translations while preserving sign language semantics. DiffSLT transforms random noise into the target latent representation, conditioned on the visual features of input video. To enhance visual conditioning, we design Guidance Fusion Module, which fully utilizes the multi-level spatiotemporal information of the visual features. We also introduce DiffSLT-P, a DiffSLT variant that conditions on pseudo-glosses and visual features, providing key textual guidance and reducing the modality gap. As a result, DiffSLT and DiffSLT-P significantly improve diversity over previous gloss-free SLT methods and achieve state-of-the-art performance on two SLT datasets, thereby markedly improving translation quality.
doi_str_mv 10.48550/arxiv.2411.17248
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2411_17248</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2411_17248</sourcerecordid><originalsourceid>FETCH-arxiv_primary_2411_172483</originalsourceid><addsrcrecordid>eNpjYJA0NNAzsTA1NdBPLKrILNMzMjE01DM0NzKx4GTwdclMSwv2CbFScM3LSMxLzsxLV3DJLEstKs4sqVTIzFMIzkzPU_BJzEsvTUxPVQgpSswrzkksyczPUyjLTFQA6S4tBvF881NSc3gYWNMSc4pTeaE0N4O8m2uIs4cu2OL4gqLM3MSiyniQA-LBDjAmrAIA1uU64g</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>DiffSLT: Enhancing Diversity in Sign Language Translation via Diffusion Model</title><source>arXiv.org</source><creator>Moon, JiHwan ; Park, Jihoon ; Kim, Jungeun ; Bae, Jongseong ; Jeon, Hyeongwoo ; Kim, Ha Young</creator><creatorcontrib>Moon, JiHwan ; Park, Jihoon ; Kim, Jungeun ; Bae, Jongseong ; Jeon, Hyeongwoo ; Kim, Ha Young</creatorcontrib><description>Sign language translation (SLT) is challenging, as it involves converting sign language videos into natural language. Previous studies have prioritized accuracy over diversity. However, diversity is crucial for handling lexical and syntactic ambiguities in machine translation, suggesting it could similarly benefit SLT. In this work, we propose DiffSLT, a novel gloss-free SLT framework that leverages a diffusion model, enabling diverse translations while preserving sign language semantics. DiffSLT transforms random noise into the target latent representation, conditioned on the visual features of input video. To enhance visual conditioning, we design Guidance Fusion Module, which fully utilizes the multi-level spatiotemporal information of the visual features. We also introduce DiffSLT-P, a DiffSLT variant that conditions on pseudo-glosses and visual features, providing key textual guidance and reducing the modality gap. As a result, DiffSLT and DiffSLT-P significantly improve diversity over previous gloss-free SLT methods and achieve state-of-the-art performance on two SLT datasets, thereby markedly improving translation quality.</description><identifier>DOI: 10.48550/arxiv.2411.17248</identifier><language>eng</language><subject>Computer Science - Computer Vision and Pattern Recognition</subject><creationdate>2024-11</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,777,882</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2411.17248$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2411.17248$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Moon, JiHwan</creatorcontrib><creatorcontrib>Park, Jihoon</creatorcontrib><creatorcontrib>Kim, Jungeun</creatorcontrib><creatorcontrib>Bae, Jongseong</creatorcontrib><creatorcontrib>Jeon, Hyeongwoo</creatorcontrib><creatorcontrib>Kim, Ha Young</creatorcontrib><title>DiffSLT: Enhancing Diversity in Sign Language Translation via Diffusion Model</title><description>Sign language translation (SLT) is challenging, as it involves converting sign language videos into natural language. Previous studies have prioritized accuracy over diversity. However, diversity is crucial for handling lexical and syntactic ambiguities in machine translation, suggesting it could similarly benefit SLT. In this work, we propose DiffSLT, a novel gloss-free SLT framework that leverages a diffusion model, enabling diverse translations while preserving sign language semantics. DiffSLT transforms random noise into the target latent representation, conditioned on the visual features of input video. To enhance visual conditioning, we design Guidance Fusion Module, which fully utilizes the multi-level spatiotemporal information of the visual features. We also introduce DiffSLT-P, a DiffSLT variant that conditions on pseudo-glosses and visual features, providing key textual guidance and reducing the modality gap. As a result, DiffSLT and DiffSLT-P significantly improve diversity over previous gloss-free SLT methods and achieve state-of-the-art performance on two SLT datasets, thereby markedly improving translation quality.</description><subject>Computer Science - Computer Vision and Pattern Recognition</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNpjYJA0NNAzsTA1NdBPLKrILNMzMjE01DM0NzKx4GTwdclMSwv2CbFScM3LSMxLzsxLV3DJLEstKs4sqVTIzFMIzkzPU_BJzEsvTUxPVQgpSswrzkksyczPUyjLTFQA6S4tBvF881NSc3gYWNMSc4pTeaE0N4O8m2uIs4cu2OL4gqLM3MSiyniQA-LBDjAmrAIA1uU64g</recordid><startdate>20241126</startdate><enddate>20241126</enddate><creator>Moon, JiHwan</creator><creator>Park, Jihoon</creator><creator>Kim, Jungeun</creator><creator>Bae, Jongseong</creator><creator>Jeon, Hyeongwoo</creator><creator>Kim, Ha Young</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20241126</creationdate><title>DiffSLT: Enhancing Diversity in Sign Language Translation via Diffusion Model</title><author>Moon, JiHwan ; Park, Jihoon ; Kim, Jungeun ; Bae, Jongseong ; Jeon, Hyeongwoo ; Kim, Ha Young</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-arxiv_primary_2411_172483</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Computer Science - Computer Vision and Pattern Recognition</topic><toplevel>online_resources</toplevel><creatorcontrib>Moon, JiHwan</creatorcontrib><creatorcontrib>Park, Jihoon</creatorcontrib><creatorcontrib>Kim, Jungeun</creatorcontrib><creatorcontrib>Bae, Jongseong</creatorcontrib><creatorcontrib>Jeon, Hyeongwoo</creatorcontrib><creatorcontrib>Kim, Ha Young</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Moon, JiHwan</au><au>Park, Jihoon</au><au>Kim, Jungeun</au><au>Bae, Jongseong</au><au>Jeon, Hyeongwoo</au><au>Kim, Ha Young</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>DiffSLT: Enhancing Diversity in Sign Language Translation via Diffusion Model</atitle><date>2024-11-26</date><risdate>2024</risdate><abstract>Sign language translation (SLT) is challenging, as it involves converting sign language videos into natural language. Previous studies have prioritized accuracy over diversity. However, diversity is crucial for handling lexical and syntactic ambiguities in machine translation, suggesting it could similarly benefit SLT. In this work, we propose DiffSLT, a novel gloss-free SLT framework that leverages a diffusion model, enabling diverse translations while preserving sign language semantics. DiffSLT transforms random noise into the target latent representation, conditioned on the visual features of input video. To enhance visual conditioning, we design Guidance Fusion Module, which fully utilizes the multi-level spatiotemporal information of the visual features. We also introduce DiffSLT-P, a DiffSLT variant that conditions on pseudo-glosses and visual features, providing key textual guidance and reducing the modality gap. As a result, DiffSLT and DiffSLT-P significantly improve diversity over previous gloss-free SLT methods and achieve state-of-the-art performance on two SLT datasets, thereby markedly improving translation quality.</abstract><doi>10.48550/arxiv.2411.17248</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2411.17248
ispartof
issn
language eng
recordid cdi_arxiv_primary_2411_17248
source arXiv.org
subjects Computer Science - Computer Vision and Pattern Recognition
title DiffSLT: Enhancing Diversity in Sign Language Translation via Diffusion Model
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-18T14%3A44%3A59IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=DiffSLT:%20Enhancing%20Diversity%20in%20Sign%20Language%20Translation%20via%20Diffusion%20Model&rft.au=Moon,%20JiHwan&rft.date=2024-11-26&rft_id=info:doi/10.48550/arxiv.2411.17248&rft_dat=%3Carxiv_GOX%3E2411_17248%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true