arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2103.05271cs.CV
收藏

场景图生成中语义歧义的概率建模

Probabilistic Modeling of Semantic Ambiguity for Scene Graph Generation

Tsinghua University(清华大学);University of Electronic Science and Technology of China(电子科技大学);Tencent AI Lab(腾讯AI实验室);The Chinese University of Hong Kong, Shenzhen(香港中文大学(深圳));Shenzhen Research Institute of Big Data(深圳大数据研究院)

Gengcong Yang, Jingyi Zhang, Yong Zhang, Baoyuan Wu, Yujiu Yang

摘要:

为了生成“准确”的场景图,几乎所有现有的方法都以确定性的方式预测成对关系。然而,我们认为,视觉关系往往是语义模糊的。具体来说,受语言学知识的启发,我们将歧义分为三种类型:同义歧义、上下义歧义和多视角歧义。这种模糊性自然会导致\emph{implicit multi-label}的问题,从而激发了对不同预测的需求。在这项工作中,我们提出了一种新的即插即用的概率不确定性建模(PUM)模块。它将每个联合区域建模为高斯分布,其方差度量相应视觉内容的不确定性。与传统的确定性方法相比,这种不确定性建模带来了特征表示的随机性,这自然使不同的预测成为可能。作为一个副产品,PUM还设法覆盖更细粒度的关系,从而解决了对频繁关系的偏见问题。在大规模Visual Genome基准测试上的大量实验表明,将PUM与新提出的ResCAGCN相结合可以实现最先进的性能,特别是在平均召回度量下。此外,我们证明了PUM的普遍有效性插入到一些现有的模型,并提供其能力,以产生不同的,但似乎合理的视觉关系有见地的分析。

英文摘要:

To generate "accurate" scene graphs, almost all existing methods predict pairwise relationships in a deterministic manner. However, we argue that visual relationships are often semantically ambiguous. Specifically, inspired by linguistic knowledge, we classify the ambiguity into three types: Synonymy Ambiguity, Hyponymy Ambiguity, and Multi-view Ambiguity. The ambiguity naturally leads to the issue of \emph{implicit multi-label}, motivating the need for diverse predictions. In this work, we propose a novel plug-and-play Probabilistic Uncertainty Modeling (PUM) module. It models each union region as a Gaussian distribution, whose variance measures the uncertainty of the corresponding visual content. Compared to the conventional deterministic methods, such uncertainty modeling brings stochasticity of feature representation, which naturally enables diverse predictions. As a byproduct, PUM also manages to cover more fine-grained relationships and thus alleviates the issue of bias towards frequent relationships. Extensive experiments on the large-scale Visual Genome benchmark show that combining PUM with newly proposed ResCAGCN can achieve state-of-the-art performances, especially under the mean recall metric. Furthermore, we prove the universal effectiveness of PUM by plugging it into some existing models and provide insightful analysis of its ability to generate diverse yet plausible visual relationships.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑