Grid R-CNN

CVPR 2019 Xin LuBuyu LiYuxin YueQuanquan LiJunjie 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... (read more)

PDF Abstract

Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT BENCHMARK
Object Detection COCO minival Grid R-CNN (ResNet-50-FPN) box AP 39.6 # 43
AP50 58.3 # 35
AP75 42.4 # 31
APS 22.6 # 29
APM 43.8 # 25
APL 51.5 # 33
Object Detection COCO minival Grid R-CNN (ResNet-101-FPN) box AP 41.3 # 36
AP50 60.3 # 26
AP75 44.4 # 25
APS 23.4 # 27
APM 45.8 # 19
APL 54.1 # 26
Object Detection COCO test-dev Grid R-CNN (ResNeXt-101-FPN) box AP 43.2 # 39
AP50 63.0 # 43
AP75 46.6 # 43
APS 25.1 # 45
APM 46.5 # 42
APL 55.2 # 41

Methods used in the Paper