Joint COCO and Mapillary Workshop at ICCV 2019: COCO Instance Segmentation Challenge Track
2020-10-06Unverified0· sign in to hype
Zeming Li, Yuchen Ma, Yukang Chen, Xiangyu Zhang, Jian Sun
Unverified — Be the first to reproduce this paper.
ReproduceAbstract
In this report, we present our object detection/instance segmentation system, MegDetV2, which works in a two-pass fashion, first to detect instances then to obtain segmentation. Our baseline detector is mainly built on a new designed RPN, called RPN++. On the COCO-2019 detection/instance-segmentation test-dev dataset, our system achieves 61.0/53.1 mAP, which surpassed our 2018 winning results by 5.0/4.2 respectively. We achieve the best results in COCO Challenge 2019 and 2020.