Strategies for training deep learning models in medical domains with small reference datasets

Mathematics

• Training 100%
• Learning 85%
• Strategy 66%
• Segmentation 42%
• Slice 38%
• Graph Cuts 36%
• Model 31%
• Demand 23%
• Transfer Learning 22%
• User Interaction 20%
• Reinforcement Learning 17%
• Acquisition 15%
• Post-processing 14%
• Cascade 12%
• Truth 12%
• Adjustment 11%
• Arbitrary 6%

Engineering & Materials Science

• Deep learning 84%
• Magnetic resonance imaging 49%
• Transfer learning 28%
• Reinforcement learning 23%