Research

Self-Erasing Network for Integral Object Attention

Visual results on a complex image with 4 semantic categories

 

 

Code and results on VOC images are available [here]

 

Per-category IoU scores are available:

VOC val set

categoryIoU (VGG)IoU (ResNet-101)
background0.8900.897
aeroplane0.8240.858
bicycle0.3240.437
bird0.7150.775
boat0.6340.617
bottle0.6270.681
bus0.8030.760
car0.7280.668
cat0.7990.832
chair0.1780.173
cow0.6920.790
dining table0.1940.100
dog0.7360.805
horse0.7180.811
motorbike0.6930.744
person0.6830.757
potted plant0.3550.389
sheep0.7300.774
sofa0.2660.196
train0.7350.676
tv/monitor0.5020.507
mean0.6110.631

 

If you think this work is helpful for you, please cite:

@inproceedings{hou2018selferasing,
  title={Self-Erasing Network for Integral Object Attention},
  author={Hou, Qibin and Jiang, Peng-Tao and Wei, Yunchao and Cheng, Ming-Ming},
  booktitle={NeurIPS},
  year={2018}
}

 

(Visited 4,306 times, 1 visits today)
Subscribe
Notify of
guest

0 Comments
Inline Feedbacks
View all comments