arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2007.08728cs.CVcs.LG
收藏

基于动作共现先验的人-物交互检测

Detecting Human-Object Interactions with Action Co-occurrence Priors

KAIST(韩国科学技术院);Microsoft Research(微软研究院)

Dong-Jin Kim, Xiao Sun, Jinsoo Choi, Stephen Lin, In So Kweon

摘要:

人-物交互(HOI)检测任务中的一个常见问题是,大量的HOI类只有少量的标记示例,导致训练集具有长尾分布。缺乏正标签可能导致这些类别的分类准确度较低。为了解决这个问题,我们观察到人与物体之间的相互作用存在自然的相关性和反相关性。在本文中,我们将相关性建模为动作同现矩阵,并提出了学习这些先验知识的技术,并利用它们进行更有效的训练,特别是在罕见的类别中。实验证明了我们方法的实用性,在两个领先的HOI检测基准数据集HICO-Det和V-COCO上,我们方法的性能超过了最先进的方法。

英文摘要:

A common problem in human-object interaction (HOI) detection task is that numerous HOI classes have only a small number of labeled examples, resulting in training sets with a long-tailed distribution. The lack of positive labels can lead to low classification accuracy for these classes. Towards addressing this issue, we observe that there exist natural correlations and anti-correlations among human-object interactions. In this paper, we model the correlations as action co-occurrence matrices and present techniques to learn these priors and leverage them for more effective training, especially in rare classes. The utility of our approach is demonstrated experimentally, where the performance of our approach exceeds the state-of-the-art methods on both of the two leading HOI detection benchmark datasets, HICO-Det and V-COCO.

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。

↑