BUSSARD：面向双射通用场景特定异常关系检测的归一化流
BUSSARD: Normalizing Flows for Bijective Universal Scene-Specific Anomalous Relationship Detection
把场景图里的物体-关系-物体三元组用语言模型嵌入后接归一化流做似然式异常检测，为视觉关系理解提供了一个可复用的『关系级异常打分』范式，可直接迁移到场景图生成的关系合理性校验或开放世界关系检测。
原文依据 A normalizing flow model is used to learn bijective transformations that map object-relation-object triplets from scene graphs to a simple base distribution
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论我们提出双射通用场景特定异常关系检测（BUSSARD），这是一种基于归一化流的模型，用于检测由图像生成的场景图中的异常关系。我们的工作采用多模态方法，利用语言模型对场景图中的物体词元与关系词元进行嵌入，以利用现实世界的语义知识。归一化流模型用于学习双射变换，将场景图中的物体-关系-物体三元组映射到简单的基础分布（通常为高斯分布），从而通过似然估计实现异常检测。我们在包含办公室与餐厅场景的 SARD 数据集上评估了该方法。与当前最优模型相比，我们的方法在 AUROC 上约提升 10%，同时速度提升五倍。通过消融实验，我们展示了更优的鲁棒性与通用性，尤其是在同义词使用方面：我们的模型保持稳定性能，而基线出现 17.5% 的偏差。这项工作展示了基于学习的方法在场景图关系异常检测方面的强大潜力。我们的代码可在 https://github.com/mschween/BUSSARD 获取。
英文原文摘要 1232 字符
We propose Bijective Universal Scene-Specific Anomalous Relationship Detection (BUSSARD), a normalizing flow-based model for detecting anomalous relations in scene graphs, generated from images. Our work follows a multimodal approach, embedding object and relationship tokens from scene graphs with a language model to leverage semantic knowledge from the real world. A normalizing flow model is used to learn bijective transformations that map object-relation-object triplets from scene graphs to a simple base distribution (typically Gaussian), allowing anomaly detection through likelihood estimation. We evaluate our approach on the SARD dataset containing office and dining room scenes. Our method achieves around 10% better AUROC results compared to the current state-of-the-art model, while simultaneously being five times faster. Through ablation studies, we demonstrate superior robustness and universality, particularly regarding the use of synonyms, with our model maintaining stable performance while the baseline shows 17.5% deviation. This work demonstrates the strong potential of learning-based methods for relationship anomaly detection in scene graphs. Our code is available at https://github.com/mschween/BUSSARD.