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

Evaluation Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK COMPARE
Object Detection COCO minival Grid R-CNN (ResNet-101-FPN) box AP 41.3 # 30
Object Detection COCO minival Grid R-CNN (ResNet-101-FPN) AP50 60.3 # 15
Object Detection COCO minival Grid R-CNN (ResNet-101-FPN) AP75 44.4 # 16
Object Detection COCO minival Grid R-CNN (ResNet-101-FPN) APS 23.4 # 22
Object Detection COCO minival Grid R-CNN (ResNet-101-FPN) APM 45.8 # 16
Object Detection COCO minival Grid R-CNN (ResNet-101-FPN) APL 54.1 # 18
Object Detection COCO minival Grid R-CNN (ResNet-50-FPN) box AP 39.6 # 36
Object Detection COCO minival Grid R-CNN (ResNet-50-FPN) AP50 58.3 # 26
Object Detection COCO minival Grid R-CNN (ResNet-50-FPN) AP75 42.4 # 22
Object Detection COCO minival Grid R-CNN (ResNet-50-FPN) APS 22.6 # 24
Object Detection COCO minival Grid R-CNN (ResNet-50-FPN) APM 43.8 # 22
Object Detection COCO minival Grid R-CNN (ResNet-50-FPN) APL 51.5 # 24
Object Detection COCO test-dev Grid R-CNN (ResNeXt-101-FPN) box AP 43.2 # 28
Object Detection COCO test-dev Grid R-CNN (ResNeXt-101-FPN) AP50 63.0 # 27
Object Detection COCO test-dev Grid R-CNN (ResNeXt-101-FPN) AP75 46.6 # 23
Object Detection COCO test-dev Grid R-CNN (ResNeXt-101-FPN) APS 25.1 # 27
Object Detection COCO test-dev Grid R-CNN (ResNeXt-101-FPN) APM 46.5 # 22
Object Detection COCO test-dev Grid R-CNN (ResNeXt-101-FPN) APL 55.2 # 22