Pre- and Post-processing Strategies for Generic Slice-wise Segmentation of Tomographic 3D datasets Utilizing U-Net Deep Learning Models Trained for Specific Diagnostic Domains

Keyphrases

• Preprocessing Techniques 100%
• Deep Learning Model 100%
• Post-processing Techniques 100%
• U-Net 100%
• 3D Data Model 100%
• Medical Image Segmentation 66%
• Liver 33%
• Memory Consumption 33%
• Deep Learning Algorithm 33%
• Training Time 33%
• Fully Automated 33%
• Training Data 33%
• Segmentation Model 33%
• Short Training 33%
• 2D Slices 33%
• Rack 33%
• TensorFlow 33%
• AI Methods 33%
• Keras 33%
• Deep Learning Libraries 33%
• Reduced Memory 33%
• Medical Image Processing 33%
• Five-layer 33%
• Semi-automated Tool 33%
• Segmentation Accuracy 33%
• Coronal View 33%
• Surface Metrics 33%
• Sagittal View 33%
• Decathlon 33%
• Scalable Deep Learning 33%
• Segmented Volumes 33%

Computer Science

• Postprocessing 100%
• Deep Learning Model 100%
• U-Net 100%
• Preprocessing 100%
• Processing Strategy 100%
• Deep Learning Method 66%
• Memory Consumption 33%
• Training Data 33%
• Segmentation Model 33%
• Keras 33%
• Adjacent Slice 33%
• segmentation accuracy 33%
• Automated Tool 33%
• Medical Image Processing 33%
• Artificial Intelligence 33%
• Graphics Processing Unit 33%

Engineering

• Processing Strategy 100%
• Deep Learning Method 100%
• Image Processing 33%
• Metrics 33%
• Similarities 33%
• Graphics Processing Unit 33%
• Medical Image 33%
• Hybrid Layer 33%
• Artificial Intelligence 33%