Vision-Language Grounding as Bidirectional Concept Correspondence

  • 类型:arxiv
  • 标识:2608.07886
  • 链接:https://arxiv.org/abs/2608.07886
  • 主分类:multimodal
  • 形态:method
  • 被引:0
  • 被引来源:Semantic Scholar + OpenAlex
  • S2被引:0
  • OpenAlex被引:0
  • 影响力被引:0
  • TLDR:This formulation unifies common grounding tasks, including phrase grounding, referring expression grounding, and open-vocabulary detection, by treating text segmentation, image segmentation, and cross-modal alignment as a single correspondence prediction problem.
  • OpenAlex ID:W7202198331
  • OpenAlex DOI:10.48550/arxiv.2608.07886
  • DOI:10.48550/arxiv.2608.07886
  • DOI来源:OpenAlex
  • 开放获取:green
  • 开放获取链接:https://arxiv.org/pdf/2608.07886
  • OpenAlex更新:2026-09-01
  • 待LLM分类:否
  • 标题中文:视觉-语言接地作为双向概念对应
  • TLDR中文:该形式化方法将短语定位、指代表达定位以及开放词汇检测等常见 grounding 任务统一起来,将文本分割、图像分割以及跨模态对齐视为单一的对应预测问题。
  • 来源文件:
  • /inbox/tom/_candidates/2026-08-11-agent-rag-longcontext-candidates.json
  • [S2 enrich]
  • [OpenAlex backfill]