Krishna Kumar Singh, Santosh Divvala, Ali Farhadi, Yong Jae Lee
TL;DR本文介绍了一种从源类别向目标类别传递常识知识以检测物体的可拓展方法,并将该方法应用于 MS COCO 数据集中,发现常识知识可以显著提高检测性能。
Abstract
We present a scalable approach for detecting objects by transferring
common-sense knowledge (DOCK) from source to target categories. In our setting,
the training data for the source categories have bounding box a