RegFormer：面向高效弱监督人-物交互检测的可迁移关系视觉定位
RegFormer: Transferable Relational Grounding for Efficient Weakly-Supervised Human-Object Interaction Detection
RegFormer 用空间视觉定位信号作为弱监督 HOI 推理的引导，把图像级交互推理直接迁移到实例级，无需额外训练；对做 HOI、场景图或弱监督定位的研究者，这是一个可复用的「定位引导关系推理」模块，能绕开枚举实例对带来的计算开销与假阳性。
原文依据 RegFormer leverages spatially grounded signals as guidance for the reasoning process and promotes locality-aware interaction learning.
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论弱监督人-物交互（HOI）检测仅依赖图像级标注学习交互，对可扩展的场景理解至关重要。由于缺乏定位信号，已有工作通常依赖外部目标检测器生成候选对，再通过成对推理推断其交互关系。然而，该框架需要枚举大量实例对，计算开销巨大，因而难以扩展。此外，它还受到非交互组合所产生假阳性的困扰，阻碍了准确的实例级 HOI 推理。为解决这些问题，我们提出关系视觉定位 Transformer（RegFormer），一个用于高效、准确 HOI 推理的通用交互识别模块。在图像级监督下，RegFormer 利用空间视觉定位信号作为推理过程的引导，并促进局部感知的交互学习。通过学习局部化的交互线索，该模块能够区分人、物体及其交互，从而无需额外训练即可从图像级交互推理直接迁移到精确且高效的实例级推理。大量实验与分析表明，RegFormer 能有效学习用于实例级交互推理的空间线索，运行效率高，甚至达到与全监督模型相当的性能。代码已发布于 https://github.com/mlvlab/RegFormer。
英文原文摘要 1521 字符
Weakly-supervised Human-Object Interaction (HOI) detection is essential for scalable scene understanding, as it learns interactions from only image-level annotations. Due to the lack of localization signals, prior works typically rely on an external object detector to generate candidate pairs and then infer their interactions through pairwise reasoning. However, this framework often struggles to scale due to the substantial computational cost incurred by enumerating numerous instance pairs. In addition, it suffers from false positives arising from non-interactive combinations, which hinder accurate instance-level HOI reasoning. To address these issues, we introduce Relational Grounding Transformer (RegFormer), a versatile interaction recognition module for efficient and accurate HOI reasoning. Under image-level supervision, RegFormer leverages spatially grounded signals as guidance for the reasoning process and promotes locality-aware interaction learning. By learning localized interaction cues, our module distinguishes humans, objects, and their interactions, enabling direct transfer from image-level interaction reasoning to precise and efficient instance-level reasoning without additional training. Our extensive experiments and analyses demonstrate that RegFormer effectively learns spatial cues for instance-level interaction reasoning, operates with high efficiency, and even achieves performance comparable to fully supervised models. Our code is available at https://github.com/mlvlab/RegFormer.