让大模型更懂视频中物体的时间位置，提升跨场景泛化能力。
SlotVTG: Object-Centric Adapter for Generalizable Video Temporal Grounding
- 用轻量级插件将视觉特征拆成语义连贯的物体槽
- 在多个数据集上显著提升跨域泛化性能，仅增加少量计算开销
- 适合需要模型适应新场景的视频理解任务
多模态大模型在视频时间定位任务中表现优异，但其粗粒度识别能力难以支持细粒度时序理解，因此任务特定微调必不可少。然而，这种微调导致模型记忆数据集特有捷径而非真实视觉内容，造成跨域（OOD）泛化能力差。基于物体中心的学习通过将场景分解为实体级表示提供了潜在解决方案，但现有方法需从头重新训练整个多阶段流程。我们提出 SlotVTG 框架，以极低代价引导大模型进行物体中心、输入驱动的视觉推理。SlotVTG 引入轻量级槽适配器，通过槽注意力将视觉标记分解为抽象槽并重构原序列，自监督视觉模型提供的物体存在先验促使语义一致的槽形成。在标准 VTG 基准上的跨域评估表明，该方法显著提升 OOD 稳健性，同时保持与基线相当的域内（ID）性能，且开销极小。
原文摘要 · Abstract (English)
Multimodal Large Language Models (MLLMs) have shown strong performance on Video Temporal Grounding (VTG). However, their coarse recognition capabilities are insufficient for fine-grained temporal understanding, making task-specific fine-tuning indispensable. This fine-tuning causes models to memorize dataset-specific shortcuts rather than faithfully grounding in the actual visual content, leading to poor Out-of-Domain (OOD) generalization. Object-centric learning offers a promising remedy by decomposing scenes into entity-level representations, but existing approaches require re-running the entire multi-stage training pipeline from scratch. We propose SlotVTG, a framework that steers MLLMs toward object-centric, input-grounded visual reasoning at minimal cost. SlotVTG introduces a lightweight slot adapter that decomposes visual tokens into abstract slots via slot attention and reconstructs the original sequence, where objectness priors from a self-supervised vision model encourage semantically coherent slot formation. Cross-domain evaluation on standard VTG benchmarks demonstrates that our approach significantly improves OOD robustness while maintaining competitive In-Domain (ID) performance with minimal overhead.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。