Fully Convolutional Networks for Semantic Segmentation

CVPR 2015 Jonathan LongEvan ShelhamerTrevor Darrell

Convolutional networks are powerful visual models that yield hierarchies of features. We show that convolutional networks by themselves, trained end-to-end, pixels-to-pixels, exceed the state-of-the-art in semantic segmentation... (read more)

PDF Abstract

Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT LEADERBOARD
Semantic Segmentation ADE20K FCN Validation mIoU 29.39 # 18
Semantic Segmentation COCO-Stuff test FCN (VGG-16) mIoU 22.7% # 9
Semantic Segmentation PASCAL VOC 2012 test FCN (VGG-16) Mean IoU 62.2% # 39
Semantic Segmentation SkyScapes-Dense FCN8s (ResNet-50) Mean IoU 33.06 # 3
Semantic Segmentation SkyScapes-Lane FCN8s (ResNet-50) Mean IoU 13.74 # 2

Results from Other Papers


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK SOURCE PAPER COMPARE
Optic Disc Segmentation Drishti-GS FCN DiceOC 0.8795 # 6
DiceOD 0.9569 # 6
mIoU 0.8392 # 5
Optic Disc Segmentation REFUGE FCN DiceOC 0.8467 # 6
DiceOD 92.56 # 6
mIoU 0.8247 # 5
Multi-tissue Nucleus Segmentation Kumar FCN8 (e) Dice 0.797 # 10
Hausdorff Distance (mm) 31.2 # 16
Semantic Segmentation PASCAL Context FCN-8s mIoU 37.8 # 29

Methods used in the Paper