SEMANTIC SEGMENTATION of AERIAL IMAGERY VIA MULTI-SCALE SHUFFLING CONVOLUTIONAL NEURAL NETWORKS with DEEP SUPERVISION

Keyphrases

• Convolutional Neural Network 100%
• Aerial Imagery 100%
• Deeply-supervised 30%
• Semantic Segmentation 20%
• Digital Surface Model 20%
• Geometric Features 20%
• Deep Supervision 20%
• Radiometric Characteristics 20%
• True Orthophoto 20%
• Benchmark Dataset 10%
• Segmentation Problem 10%
• Performance Evaluation 10%
• Numerical Results 10%
• Feature Map 10%
• Multimodal Data 10%
• Upsampling 10%
• Multi-scale Features 10%
• Scale Extension 10%
• Additional Losses 10%
• Improved Detection 10%
• Shuffle Operator 10%
• Multi-scale Fusion 10%

Computer Science

• Convolutional Neural Network 100%
• Image Segmentation 20%
• Segmentation Task 10%
• Performance Evaluation 10%
• Feature Map 10%