fcn

Introduction

Official Repo

Code Snippet

FCN (CVPR'2015/TPAMI'2017)

@inproceedings{long2015fully,
    title={Fully convolutional networks for semantic segmentation},
    author={Long, Jonathan and Shelhamer, Evan and Darrell, Trevor},
    booktitle={Proceedings of the IEEE conference on computer vision and pattern recognition},
    pages={3431--3440},
    year={2015}
}

Results

PASCAL VOC

Backbone	Pretrain	Crop Size	Schedule	Train/Eval Set	mIoU	Download
R-50-D8	ImageNet-1k-224x224	512x512	LR/POLICY/BS/EPOCH: 0.01/poly/16/60	trainaug/val	67.80%	cfg | model | log
R-50-D16	ImageNet-1k-224x224	512x512	LR/POLICY/BS/EPOCH: 0.01/poly/16/60	trainaug/val	66.58%	cfg | model | log
R-101-D8	ImageNet-1k-224x224	512x512	LR/POLICY/BS/EPOCH: 0.01/poly/16/60	trainaug/val	70.59%	cfg | model | log
R-101-D16	ImageNet-1k-224x224	512x512	LR/POLICY/BS/EPOCH: 0.01/poly/16/60	trainaug/val	69.39%	cfg | model | log

ADE20k

Backbone	Pretrain	Crop Size	Schedule	Train/Eval Set	mIoU	Download
R-50-D8	ImageNet-1k-224x224	512x512	LR/POLICY/BS/EPOCH: 0.01/poly/16/130	train/val	36.96%	cfg | model | log
R-50-D16	ImageNet-1k-224x224	512x512	LR/POLICY/BS/EPOCH: 0.01/poly/16/130	train/val	36.50%	cfg | model | log
R-101-D8	ImageNet-1k-224x224	512x512	LR/POLICY/BS/EPOCH: 0.01/poly/16/130	train/val	41.22%	cfg | model | log
R-101-D16	ImageNet-1k-224x224	512x512	LR/POLICY/BS/EPOCH: 0.01/poly/16/130	train/val	40.15%	cfg | model | log

CityScapes

Backbone	Pretrain	Crop Size	Schedule	Train/Eval Set	mIoU	Download
R-50-D8	ImageNet-1k-224x224	512x1024	LR/POLICY/BS/EPOCH: 0.01/poly/8/220	train/val	75.16%	cfg | model | log
R-50-D16	ImageNet-1k-224x224	512x1024	LR/POLICY/BS/EPOCH: 0.01/poly/8/220	train/val	73.94%	cfg | model | log
R-101-D8	ImageNet-1k-224x224	512x1024	LR/POLICY/BS/EPOCH: 0.01/poly/8/220	train/val	76.31%	cfg | model | log
R-101-D16	ImageNet-1k-224x224	512x1024	LR/POLICY/BS/EPOCH: 0.01/poly/8/220	train/val	75.36%	cfg | model | log

More

You can also download the model weights from following sources:

• BaiduNetdisk: https://pan.baidu.com/s/1gD-NJJWOtaHCtB0qHE79rA with access code s757