Comparative research on different backbone architectures of DeepLabV3+ for building segmentation

Keyphrases

• DeepLabv3+ 100%
• Building Segmentation 100%
• Backbone Design 100%
• INRIA 66%
• Building Extraction 66%
• ResNet50 Model 66%
• Land Use 33%
• Time Consumption 33%
• Remote Sensing Image 33%
• Deep Neural Network 33%
• Evaluation Metrics 33%
• F1 Score 33%
• Aerial Image 33%
• Training Time 33%
• Public Dataset 33%
• Man-made 33%
• Testing Time 33%
• High Resolution Satellite Images 33%
• Massachusetts 33%
• Image Labeling 33%
• Rapid Urban Growth 33%
• Massachusetts Building Dataset 33%
• Convolutional Neural Network Architecture 33%
• High-resolution Aerial Images 33%
• Automatic Building Extraction 33%
• MobileNetV2 33%
• Xception Model 33%
• University Buildings 33%
• Image Sources 33%
• Evaluation Training 33%
• Urban Globalization 33%
• Wuhan University 33%
• Building Classes 33%
• Automatic Interpretation 33%
• Remote Sensing Interpretation 33%
• Earth Observation Missions 33%
• Intersection over Union 33%
• ResNet18 33%

Earth and Planetary Sciences

• Land Use 100%
• Remote Sensing 100%
• Satellite Image 100%
• Urban Growth 100%
• Testing Time 100%