RAO-UNet: a residual attention and octave UNet for road crack detection via balance loss

Computer Science

• Frequency Feature 100%
• Detection Accuracy 50%
• Computational Cost 50%
• Process Optimization 50%
• Detection Result 50%
• Processing Speed 50%
• Spatial Frequency 50%
• Image Frequency 50%
• Public Data Set 50%
• Optimization Space 50%

Engineering

• Road 100%
• Crack Detection 100%
• Computational Cost 20%
• State-of-the-Art Method 20%
• Computational Fluid Dynamics 20%
• Generated Image 20%
• Loss Function 20%
• Image Frequency 20%
• Spatial Frequency 20%
• Missed Detection 20%
• Frequency Relationship 20%