Using ResNet3D_FeatureExtractor for Classification:

• Train: 10000 / Val: 657
• EPOCH = 2 / NUM_CLASSES = 18 / LEARNING_RATE = 0.0001 / BATCH_SIZE = 8
• Results:
  • Epoch 1 Summary | Train Loss: 0.3688 | Val Loss: 0.3322 | AUROC: 0.8520 | Accuracy: 0.1028 | Sensitivity: 0.3899 | Specificity: 0.9570
  • Epoch 2 Summary | Train Loss: 0.3383 | Val Loss: 0.3199 | AUROC: 0.8657 | Accuracy: 0.1503 | Sensitivity: 0.4198 | Specificity: 0.9541