An expert knowledge-empowered CNN approach for welding radiographic image recognition

Keyphrases

• Image Recognition 100%
• Expert Knowledge 100%
• Radiograph 100%
• Convolutional Neural Network 100%
• Neural Network Method 100%
• Self-supervised Learning 42%
• Neural Network 28%
• Learning Process 28%
• Small Sample Size 28%
• Convolutional Neural Network Model 28%
• Explainability 28%
• Non-destructive Testing 14%
• Fast Convergence 14%
• Aerospace 14%
• Learning Task 14%
• Recognition Accuracy 14%
• Extraction Efficiency 14%
• Speed-accuracy 14%
• Deep Learning Methods 14%
• Image Deblurring 14%
• Image Brightness 14%
• F1 Score 14%
• Inductive Bias 14%
• Expert Experience 14%
• Comprehensive Index 14%
• Global Aspects 14%
• Weld Defects 14%
• Adaptive Extraction 14%
• Radiographic Film 14%
• Brightness Adjustment 14%
• Decision-making Bases 14%
• Feature Adaptive 14%
• Local Aspects 14%

Computer Science

• Convolutional Neural Network 100%
• Neural Network Approach 100%
• Expert Knowledge 100%
• Self-Supervised Learning 37%
• Learning Process 25%
• Neural Network Model 25%
• Small Sample Size 25%
• Deep Learning Method 12%
• Fast Convergence 12%
• Evaluation Index 12%
• Face Recognition 12%
• Decision-Making 12%
• Recognition Accuracy 12%
• Convergence Speed 12%

Engineering

• Image Recognition 100%
• Neural Network Approach 100%
• Convolutional Neural Network 100%
• Network Model 25%
• Recognition Accuracy 12%
• Welds 12%
• Learning Task 12%
• Convergence Speed 12%
• Non-Destructive Testing 12%
• Weld Defect 12%
• Deep Learning Method 12%