Grid R-CNN

CVPR 2019  ·  Xin Lu, Buyu Li, Yuxin Yue, Quanquan Li, Junjie Yan ·

This paper proposes a novel object detection framework named Grid R-CNN, which adopts a grid guided localization mechanism for accurate object detection. Different from the traditional regression based methods, the Grid R-CNN captures the spatial information explicitly and enjoys the position sensitive property of fully convolutional architecture. Instead of using only two independent points, we design a multi-point supervision formulation to encode more clues in order to reduce the impact of inaccurate prediction of specific points. To take the full advantage of the correlation of points in a grid, we propose a two-stage information fusion strategy to fuse feature maps of neighbor grid points. The grid guided localization approach is easy to be extended to different state-of-the-art detection frameworks. Grid R-CNN leads to high quality object localization, and experiments demonstrate that it achieves a 4.1% AP gain at IoU=0.8 and a 10.0% AP gain at IoU=0.9 on COCO benchmark compared to Faster R-CNN with Res50 backbone and FPN architecture.

PDF Abstract CVPR 2019 PDF CVPR 2019 Abstract

Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Object Detection COCO minival Grid R-CNN (ResNet-101-FPN) box AP 41.3 # 151
AP50 60.3 # 75
AP75 44.4 # 66
APS 23.4 # 62
APM 45.8 # 47
APL 54.1 # 60
Object Detection COCO minival Grid R-CNN (ResNet-50-FPN) box AP 39.6 # 173
AP50 58.3 # 92
AP75 42.4 # 77
APS 22.6 # 67
APM 43.8 # 59
APL 51.5 # 71
Object Detection COCO test-dev Grid R-CNN (ResNeXt-101-FPN) box mAP 43.2 # 156
AP50 63.0 # 102
AP75 46.6 # 105
APS 25.1 # 94
APM 46.5 # 94
APL 55.2 # 99
Hardware Burden None # 1
Operations per network pass None # 1
2D Object Detection SARDet-100K Grid RCNN box mAP 48.8 # 9

Methods