Title
Recognizing novel patterns via adversarial learning for one-shot semantic segmentation.
Abstract
•Adversarial learning method is proposed for one-shot semantic segmentation.•Segmentation network can learn patterns of novel classes with only one annotation.•A discriminator differentiating the predicted map from the ground truth is designed.•We achieved the cross-validate mIoU of 49.7% on the PASCAL VOC 2012 dataset.
Year
DOI
Venue
2020
10.1016/j.ins.2020.01.016
Information Sciences
Keywords
Field
DocType
Semantic segmentation,One-shot learning,Adversarial learning,Generative adversarial networks
Discriminator,Convolutional neural network,Trustworthiness,Segmentation,One shot,Ground truth,Artificial intelligence,Pixel,Machine learning,Mathematics,Adversarial system
Journal
Volume
ISSN
Citations 
518
0020-0255
0
PageRank 
References 
Authors
0.34
0
4
Name
Order
Citations
PageRank
Guangchao Yang100.34
Dongmei Niu226.44
Caiming Zhang344688.19
Xiuyang Zhao47313.60