English(EN) Graph it first! Enabling Reasoning on Long-form Egocentric Videos through Scene Graphs

主观场景图谱赋能多模态大语言模型推理长视频

作者 PulseAugur 编辑部 · [2 个来源] · 2026-06-24 13:55

研究人员开发了一个新框架，使多模态大语言模型（MLLMs）能够推理长篇主观视频，克服了令牌限制。该方法利用主观场景图谱（EgoSGs），这是一种时间锚定、结构化的对象、属性、空间关系和交互的表示。通过将视频转换为这些紧凑的、符号化的场景图谱，该方法显著减少了输入长度，同时保留了重要的语义和时间信息，使MLLMs能够在其上下文窗口内处理整个视频序列。该技术在HD-EPIC VQA基准测试中取得了最先进的成果，优于现有的基于视频的基线。 AI

影响使MLLMs能够处理和推理扩展的视频内容，可能改进视频分析和理解方面的应用。

排序理由关于使用LLMs进行视频理解新方法的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.CV 阅读 →

AI 生成摘要 · Google Gemini · 来自 2 个来源。我们如何撰写摘要 →

报道来源 [2]

arXiv cs.CV TIER_1 English(EN) · Agnese Taluzzi, Riccardo Santambrogio, Simone Mentasti, Chiara Plizzari, Matteo Matteucci · 2026-06-25 04:00

Graph it first! Enabling Reasoning on Long-form Egocentric Videos through Scene Graphs

arXiv:2606.25842v1 Announce Type: new Abstract: Existing multi-modal large language models (MLLMs) face significant challenges in processing long video sequences due to strict input token limitations. As a result, current video understanding approaches, especially in egocentric s…
arXiv cs.CV TIER_1 English(EN) · Matteo Matteucci · 2026-06-24 13:55

Graph it first! Enabling Reasoning on Long-form Egocentric Videos through Scene Graphs

Existing multi-modal large language models (MLLMs) face significant challenges in processing long video sequences due to strict input token limitations. As a result, current video understanding approaches, especially in egocentric settings characterized by complex dynamics, frequ…

报道来源 [2]

Graph it first! Enabling Reasoning on Long-form Egocentric Videos through Scene Graphs

Graph it first! Enabling Reasoning on Long-form Egocentric Videos through Scene Graphs

相关实体

相关话题