A new weakly supervised approach for ALS point cloud semantic segmentation

Keyphrases

• Unsupervised Learning 100%
• Supervised Approach 100%
• Point Cloud Segmentation 100%
• Airborne Laser Scanning Point Clouds 100%
• Sparse Labels 66%
• Urban Areas 33%
• Computational Efficiency 33%
• Classification Performance 33%
• 3D Data 33%
• Overall Accuracy 33%
• Consistency Condition 33%
• Effective Model 33%
• Nonparametric 33%
• Entropy Regularization 33%
• Deep Learning 33%
• Labeled Data 33%
• Label Information 33%
• Benchmark Dataset 33%
• Multiple Classes 33%
• Segmentation Scheme 33%
• Unlabeled Data 33%
• Current Prediction 33%
• F1 Score 33%
• Semantic Segmentation 33%
• Implicit Information 33%
• Weakly Supervised Method 33%
• Data Annotation 33%
• Class Overlap 33%
• Incomplete Labels 33%
• ISPRS Vaihingen 33%
• Predictive Probability 33%
• Pseudo-labeling 33%
• Sparsest Point 33%
• Point Annotation 33%
• Full Supervision 33%
• Labeling Strategy 33%
• Ensemble Forecast 33%

Computer Science

• Image Segmentation 100%
• Point Cloud 100%
• Annotation 50%
• Deep Learning Method 25%
• Unlabeled Data 25%
• Intensive Task 25%
• Supervised Method 25%
• Implicit Information 25%
• Regularization 25%
• Computational Efficiency 25%
• Classification Performance 25%

Earth and Planetary Sciences

• Airborne Laser Scanning 100%
• State of the Art 33%