Deep region-based object detector consists of a region proposal step and a deep object recognition step. In this paper, we make significant improvements on both of the two steps. For region proposal we propose a novel lightweight cascade structure which can effectively improve RPN proposal quality. For object recognition we re-implement global context modeling with a few modications and obtain a performance boost (4.2% mAP gain on the ILSVRC 2016 validation set). Besides, we apply the idea of pre-training extensively and show its importance in both steps. Together with common training and testing tricks, we improve Faster R-CNN baseline by a large margin. In particular, we obtain 87.9% mAP on the PASCAL VOC 2012 test set, 65.3% on the ILSVRC 2016 test set and 36.8% on the COCO test-std set.
@article{arxiv.1710.10749,
title = {Cascade Region Proposal and Global Context for Deep Object Detection},
author = {Qiaoyong Zhong and Chao Li and Yingying Zhang and Di Xie and Shicai Yang and Shiliang Pu},
journal= {arXiv preprint arXiv:1710.10749},
year = {2017}
}