MD Loss: Efficient Training of 3-D Seismic Fault Segmentation Network Under Sparse Labels by Weakening Anomaly Annotation.
Yimin DouKewen LiJianbing ZhuTiming LiShaoquan TanZongchao HuangPublished in: IEEE Trans. Geosci. Remote. Sens. (2022)
Keyphrases
- annotation effort
- network traffic
- training set
- image labeling
- training examples
- fault detection
- anomaly detection
- segmentation algorithm
- fully automatic
- medical images
- fault diagnosis
- labelled data
- semantic segmentation
- network management
- training process
- region growing
- feature selection
- signal processing
- peer to peer
- active learning
- image segmentation
- segmentation method
- complex networks
- intrusion detection
- level set
- shape prior
- supervised learning
- wireless sensor networks
- recurrent networks
- semantic labels
- multiscale
- fault management