https://vilab.team/tag/%E5%8A%A8%E4%BD%9C%E5%BB%BA%E6%A8%A1/ 动作建模 Hugo Blox Builder (https://hugoblox.com)en-usMon, 20 Apr 2026 00:00:00 +0000 https://vilab.team/media/icon_hu2896232876136423579.png https://vilab.team/tag/%E5%8A%A8%E4%BD%9C%E5%BB%BA%E6%A8%A1/ https://vilab.team/publication/reactid-synchronizing-realistic-actions-and-identity-in-pers/ Mon, 20 Apr 2026 00:00:00 +0000 https://vilab.team/publication/reactid-synchronizing-realistic-actions-and-identity-in-pers/ <p>本文提出ReactID框架，旨在协调个性化视频生成中身份一致性与动作真实性的矛盾。针对主体-视频对齐不精确、训练不稳定、细粒度动作建模不足三大挑战，从数据、训练和动作建模三方面协同改进：构建高精度标注的ReactID-Data数据集；设计由易到难的渐进式训练课程；提出基于时间线的条件机制，通过主体感知交叉注意力和时间自适应RoPE，将子动作与特定主体绑定并嵌入时间坐标，从而生成更自然、可控的视频。</p> https://vilab.team/event/%E7%A5%9D%E8%B4%BA%E5%AE%9E%E9%AA%8C%E5%AE%A4%E7%A7%91%E7%A0%94%E6%88%90%E6%9E%9C%E5%8F%91%E8%A1%A8%E4%BA%8E-iclr/ Mon, 20 Apr 2026 00:00:00 +0000 https://vilab.team/event/%E7%A5%9D%E8%B4%BA%E5%AE%9E%E9%AA%8C%E5%AE%A4%E7%A7%91%E7%A0%94%E6%88%90%E6%9E%9C%E5%8F%91%E8%A1%A8%E4%BA%8E-iclr/ <p>热烈祝贺李蔚同学！论文《ReactID: Synchronizing Realistic Actions and Identity in Personalized Video Generation》已发表在 <em>ICLR 2026</em>。</p> <h2 id="reactid-synchronizing-realistic-actions-and-identity-in-personalized-video-generation">ReactID: Synchronizing Realistic Actions and Identity in Personalized Video Generation</h2> <p><strong>祝贺李蔚同学！该论文已发表在 <em>ICLR</em>。</strong></p> <p> <figure > <div class="d-flex justify-content-center"> <div class="w-100" ><img alt="ReactID: Synchronizing Realistic Actions and Identity in Personalized Video Generation" srcset=" /event/%E7%A5%9D%E8%B4%BA%E5%AE%9E%E9%AA%8C%E5%AE%A4%E7%A7%91%E7%A0%94%E6%88%90%E6%9E%9C%E5%8F%91%E8%A1%A8%E4%BA%8E-iclr/images/paper-01_hu10185920340004675066.webp 400w, /event/%E7%A5%9D%E8%B4%BA%E5%AE%9E%E9%AA%8C%E5%AE%A4%E7%A7%91%E7%A0%94%E6%88%90%E6%9E%9C%E5%8F%91%E8%A1%A8%E4%BA%8E-iclr/images/paper-01_hu15268550746155155384.webp 760w, /event/%E7%A5%9D%E8%B4%BA%E5%AE%9E%E9%AA%8C%E5%AE%A4%E7%A7%91%E7%A0%94%E6%88%90%E6%9E%9C%E5%8F%91%E8%A1%A8%E4%BA%8E-iclr/images/paper-01_hu4770338007721801293.webp 1200w" src="https://vilab.team/event/%E7%A5%9D%E8%B4%BA%E5%AE%9E%E9%AA%8C%E5%AE%A4%E7%A7%91%E7%A0%94%E6%88%90%E6%9E%9C%E5%8F%91%E8%A1%A8%E4%BA%8E-iclr/images/paper-01_hu10185920340004675066.webp" width="760" height="288" loading="lazy" data-zoomable /></div> </div></figure> </p> <ul> <li><strong>作者：</strong> Wei Li、Yiheng Zhang、Fuchen Long、Zhaofan Qiu、Ting Yao、Xiaoyan Sun、Tao Mei</li> <li><strong>发表载体：</strong> <em>ICLR</em></li> <li><strong>发表时间：</strong> 2026年4月20日</li> <li><strong>相关链接：</strong> <a href="https://proceedings.iclr.cc/paper_files/paper/2026/hash/6600458132d025c68a01e82081597b32-Abstract-Conference.html" target="_blank" rel="noopener">论文链接</a> · <a href="https://proceedings.iclr.cc/paper_files/paper/2026/file/6600458132d025c68a01e82081597b32-Paper-Conference.pdf" target="_blank" rel="noopener">PDF</a></li> </ul> <h3 id="论文介绍">论文介绍</h3> <p>本文提出ReactID框架，旨在协调个性化视频生成中身份一致性与动作真实性的矛盾。针对主体-视频对齐不精确、训练不稳定、细粒度动作建模不足三大挑战，从数据、训练和动作建模三方面协同改进：构建高精度标注的ReactID-Data数据集；设计由易到难的渐进式训练课程；提出基于时间线的条件机制，通过主体感知交叉注意力和时间自适应RoPE，将子动作与特定主体绑定并嵌入时间坐标，从而生成更自然、可控的视频。</p> https://vilab.team/publication/hierarchical-task-aware-temporal-modeling-and-matching-for-f/ Tue, 01 Apr 2025 00:00:00 +0000 https://vilab.team/publication/hierarchical-task-aware-temporal-modeling-and-matching-for-f/ <p>本文针对少样本动作识别中训练样本稀缺且视频结构复杂的问题，提出分层任务感知时间建模与匹配方法（HTTMM）。该方法通过分层结构充分建模时空特征，并利用任务感知机制增强对关键运动模式的感知，从而提升查询样本与支持样本之间的匹配效果。在多个基准数据集上的实验验证了其有效性，尤其适用于需要局部运动感知的细粒度动作分类任务。</p>