Relation Networks for Object Detection https://arxiv.org/abs/1711.11575
Towards High Performance Video Object Detection https://arxiv.org/abs/1711.11577
Learning to Segment Every Thing http://cn.arxiv.org/pdf/1711.10370.pdf
前面两个是jifeng dai和同事的作品，后面是ross 和kaiming和同事的作品。