论文标题
用于检测人类对象相互作用的空间启动
Spatial Priming for Detecting Human-Object Interactions
论文作者
论文摘要
人类和对象的相对空间布局是确定它们如何相互作用的重要提示。但是，到目前为止，空间布局已被用作检测人类对象相互作用（HOI）的侧信息。在本文中，我们提出了一种利用此空间布局信息来检测图像中HOI的方法。提出的方法由一个布局模块组成，该布局模块素描一个视觉模块，以预测人与对象之间的相互作用类型。视觉和布局模块通过几个阶段的横向连接共享信息。该模型将布局模块中的预测用作视觉模块之前的一个，并且从视觉模块中的预测作为最终输出。它还使用Word2Vec向量包含有关对象的语义信息。对于HICO-DET数据集，所提出的模型的地图为24.79％，该数据集的绝对点比当前的最新目标高约2.8％。
The relative spatial layout of a human and an object is an important cue for determining how they interact. However, until now, spatial layout has been used just as side-information for detecting human-object interactions (HOIs). In this paper, we present a method for exploiting this spatial layout information for detecting HOIs in images. The proposed method consists of a layout module which primes a visual module to predict the type of interaction between a human and an object. The visual and layout modules share information through lateral connections at several stages. The model uses predictions from the layout module as a prior to the visual module and the prediction from the visual module is given as the final output. It also incorporates semantic information about the object using word2vec vectors. The proposed model reaches an mAP of 24.79% for HICO-Det dataset which is about 2.8% absolute points higher than the current state-of-the-art.
