arXivDaily arXiv每日学术速递 周一至周五更新

期刊&会议

International Conference on Computer Vision · 会议 · Computer Vision

2026-05-29 至 2026-05-29 共收录 1
2507.09266 2026-05-29 cs.CV

SAGE: Segment-Aware Gloss-Free Encoding for Token-Efficient Sign Language Translation

SAGE: 面向令牌高效手语翻译的分段感知无词汇编码

JianHe Low, Ozge Mercanoglu Sincan, Richard Bowden

机构 * CVSSP, University of Surrey, United Kingdom(CVSSP,萨里大学,英国)

AI总结 提出分段感知视觉标记化框架,通过手语分段将连续视频转换为离散视觉令牌,结合令牌级对比对齐和双层监督,在减少序列长度50%的同时,在PHOENIX14T基准上超越现有方法。

Comments Accepted in International Conference on Computer Vision (ICCV) Workshops. Code released at https://github.com/JianHe0628/SAGE

详情

展开后加载摘要…

URL PDF HTML 收藏