Deep point embedding for urban classification using ALS point clouds: A new perspective from local to global

Keyphrases

• Global-local 100%
• Airborne Laser Scanning Point Clouds 100%
• Urban Classification 100%
• Classification Results 66%
• Overall Accuracy 66%
• Point Cloud 66%
• Semantic Interpretation 66%
• Semantic Label 66%
• Semantic Labeling 66%
• Optimization Problem 33%
• Optimization Approach 33%
• Classification Methods 33%
• Non-homogeneity 33%
• Global Optimization 33%
• Feature-based 33%
• Local Context 33%
• Point-based 33%
• Semantic Classes 33%
• Structure Optimization 33%
• Discriminative Features 33%
• Contextual Information 33%
• Low-dimensional Space 33%
• Benchmark Dataset 33%
• High Classification Accuracy 33%
• Nonlinear Manifold 33%
• Hierarchical Data 33%
• Graph Cuts 33%
• 3D Scene 33%
• Joint Learning 33%
• Graph Optimization 33%
• Data Augmentation 33%
• Point Cloud Dataset 33%
• Markov Random Field 33%
• Global Graph 33%
• Nonlinear Manifold Learning 33%
• Feature Dimension Reduction 33%
• ISPRS Benchmark 33%
• PointNet++ 33%
• Reduced Graph 33%
• Point Cloud Processing 33%
• Challenging Problems 33%
• Learning Methods 33%
• Large-scale Dataset 33%
• Cloud Applications 33%
• Multi-scale Deep Learning 33%
• Hierarchical Deep Features 33%

Computer Science

• Point Cloud 100%
• Deep Feature 50%
• Semantic Label 33%
• classification result 33%
• 3d Scenes 16%
• Optimization Problem 16%
• Deep Learning Method 16%
• Global Optimization 16%
• Discriminative Feature 16%
• Lower Dimensional Space 16%
• Feature Dimension 16%
• Graph Cut 16%
• Cloud Application 16%
• Data Augmentation 16%
• Context Information 16%
• Classification Accuracy 16%
• Classification Method 16%
• Manifold Learning 16%

Chemical Engineering

• Deep Learning Method 100%