Densely Connected Swin-UNet for Multiscale Information Aggregation in Medical Image Segmentation

Keyphrases

• Decoder 100%
• Densely Connected 100%
• Medical Image Segmentation 100%
• Encoder-decoder 100%
• Shifted Window 100%
• Self-attention 100%
• Swin-Unet 100%
• Information Aggregation 100%
• Multi-scale Information 100%
• Competitive Performance 50%
• Computer Vision 50%
• Window-based 50%
• Deep Learning Methods 50%
• Encoder 50%
• Semantic Features 50%
• Long-range Dependence 50%
• Feature Information 50%
• Validation Metrics 50%
• U-shape 50%
• Feature Samples 50%
• Attention-based 50%
• Feature Map 50%
• Tokenization 50%
• U-Net 50%
• Segmentation Network 50%
• Skip Connection 50%
• Comprehensive Validation 50%
• Cardiac Segmentation 50%
• Image Semantic Segmentation 50%
• Feature Image 50%
• Deep Supervision 50%
• End-to-end Convolutional Neural Network 50%
• Window Approach 50%
• Aggregated Features 50%
• Long Skip Connection 50%
• Dense Prediction 50%
• Multi-scale Feature Information 50%
• Swin Transformer 50%
• Shape Segmentation 50%

Computer Science

• Image Segmentation 100%
• Information Aggregation 100%
• Feature Information 66%
• Self-Attention 66%
• Computer Vision 33%
• Semantic Feature 33%
• Feature Map 33%
• Range Dependency 33%
• Deep Learning Technique 33%
• Coder 33%
• Convolutional Neural Network 33%
• Lexical Tokenization 33%
• Swin Transformer 33%
• Shape Segmentation 33%

Engineering

• Multiscale 100%
• Medical Image Segmentation 100%
• Metrics 33%
• Computervision 33%
• Full Scale 33%
• Convolutional Neural Network 33%
• Deep Learning Method 33%
• Learning Technique 33%
• Swin Transformer 33%
• Shape Segmentation 33%