Object-Contextual Representations for Semantic Segmentation

24 Sep 2019Yuhui YuanXilin ChenJingdong Wang

In this paper, we address the semantic segmentation problem with a focus on the context aggregation strategy. Our motivation is that the label of a pixel is the category of the object that the pixel belongs to... (read more)

PDF Abstract

Evaluation Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK USES EXTRA
TRAINING DATA
COMPARE
Semantic Segmentation ADE20K val OCR (ResNet-101) mIoU 45.28% # 3
Semantic Segmentation ADE20K val OCR (HRNetV2-W48) mIoU 45.66% # 2
Semantic Segmentation Cityscapes test OCR (ResNet-101, coarse) Mean IoU (class) 82.4% # 7
Semantic Segmentation Cityscapes test HRNetV2 + OCR (w/ ASP) Mean IoU (class) 83.7% # 2
Semantic Segmentation Cityscapes test HRNetV2 + OCR + Segfix Mean IoU (class) 84.5% # 1
Semantic Segmentation Cityscapes test OCR (ResNet-101) Mean IoU (class) 81.8% # 11
Semantic Segmentation Cityscapes test OCR (HRNetV2-W48, coarse) Mean IoU (class) 83.0% # 4
Semantic Segmentation Cityscapes val OCR (ResNet-101-FCN) mIoU 80.60% # 2
Semantic Segmentation COCO-Stuff test OCR (HRNetV2-W48) mIoU 40.5% # 1
Semantic Segmentation COCO-Stuff test OCR (ResNet-101) mIoU 39.5% # 4
Semantic Segmentation LIP val OCR (HRNetV2-W48) mIoU 56.65% # 2
Semantic Segmentation LIP val OCR (ResNet-101) mIoU 55.60% # 4
Semantic Segmentation PASCAL Context OCR (HRNetV2-W48) mIoU 56.2 # 1
Semantic Segmentation PASCAL Context OCR (ResNet-101) mIoU 54.8 # 2
Semantic Segmentation PASCAL VOC 2012 test OCR (ResNet-101) Mean IoU 84.3% # 10
Semantic Segmentation PASCAL VOC 2012 test OCR (HRNetV2-W48) Mean IoU 84.5% # 9