Keyphrases
Deep Learning Network
100%
U-Net
100%
Shield Tunnel Lining
100%
Hybrid Attention
100%
Position Attention Module
80%
Channel Attention Module
60%
Spatial Dimension
40%
Balanced Accuracy
40%
F1 Score
40%
Channel Dimensions
40%
Tunnel Lining Defect
40%
Intersection over Union
40%
Performance Improvement
20%
Channel Network
20%
Cracking Behavior
20%
Field Practice
20%
Mask R-CNN
20%
Evaluation Metrics
20%
Discontinuity Problem
20%
Shanghai Metro
20%
Attention Module
20%
Convolutional Layer
20%
Crack Image
20%
Dependency Features
20%
Crack Segmentation
20%
Global Dependencies
20%
LinkNet
20%
Hybrid Positioning
20%
Crack Discontinuity
20%
PANet
20%
PSPNet
20%
Channel-wise Feature
20%
Metro Shield Tunnel
20%
State-of-the-art Models
20%
Engineering
Deep Learning Method
100%
Tunnel
100%
Spatial Dimension
50%
Channel Dimension
50%
Interdependency
25%
Testing Dataset
25%
Metrics
25%
Metro
25%
Computer Science
Deep Learning Method
100%
U-Net
100%
Spatial Dimension
40%
Channel Network
20%
Evaluation Metric
20%
Convolution Layer
20%
Interdependency
20%
Earth and Planetary Sciences
Tunnel Lining
100%
State of the Art
33%