A Multi-branch Hybrid Transformer Network for Corneal Endothelial Cell Segmentation

Keyphrases

• Multi-branch 100%
• Hybrid Transformer 100%
• Corneal Endothelial Cell Segmentation 100%
• Transformer Network 100%
• Segmentation Method 50%
• Tremor 50%
• State-of-the-art Techniques 50%
• Transformer 50%
• Global Context 50%
• Network Applications 50%
• Receptive Field 50%
• Contextual Information 50%
• Deep Learning Segmentation 50%
• Coefficient of Variation 50%
• Position Information 50%
• Full Use 50%
• Clinical Indicators 50%
• Cell Edge 50%
• Edge Position 50%
• Detail Information 50%
• Transformer Connection 50%
• Density Coefficient 50%
• Local Texture Features 50%
• Local Consistency 50%
• Texture Feature Extraction 50%
• Local Convolution 50%
• Downsampling 50%
• Hexagonality 50%
• Residual Connection 50%
• Cell Density 50%
• Long-range Dependence 50%
• Convolutional Block 50%
• Self-collected 50%

Engineering

• Feature Extraction 100%
• Deep Learning 100%
• State-of-the-Art Method 100%
• Receptive Field 100%
• Segmentation Method 100%
• Texture Feature 100%
• Context Information 100%
• Position Information 100%
• Coefficient of Variation 100%