README: MONAI Dental X-ray Analysis Class Materials

This package contains all materials needed for teaching a class on training AI models for dental X-ray analysis using the MONAI framework. The materials are designed to be accessible to dental professionals with minimal programming experience.

• dataset_research.md: Comprehensive review of available dental X-ray datasets
• dataset_preprocessing_guide.md: Guide to labeling and preprocessing dental X-rays
• dataset_preparation.md: Step-by-step instructions for preparing datasets for MONAI

• monai_model_review.md: High-level overview of MONAI for dental students
• monai_training_script.py: Complete training script with detailed comments
• validation_report.md: Validation of the training pipeline

• visual_explanations.md: Visual guide to each step of the AI pipeline
• troubleshooting_guide.md: Solutions to common issues during training

• todo.md: Checklist of completed preparation steps

1. Review the monai_model_review.md to understand the MONAI framework
2. Ensure all prerequisites are installed (see Setup section below)
3. Test the training script on a small dataset

1. Begin with the visual explanations to introduce key concepts
2. Walk through the dataset preparation steps
3. Run the training script, explaining each component
4. Show evaluation results and discuss clinical implications
5. Address common issues using the troubleshooting guide

• Python 3.7+
• PyTorch 1.9+
• MONAI 0.9+
• CUDA-capable GPU (recommended but not required)

# Create virtual environment
python -m venv dental_ai_env
source dental_ai_env/bin/activate  # On Windows: dental_ai_env\Scripts\activate

# Install dependencies
pip install torch torchvision
pip install monai
pip install matplotlib scikit-learn pandas

# Verify installation
python -c "import monai; print(monai.__version__)"

1. Follow instructions in dataset_preparation.md to organize your dental X-ray dataset
2. Use the preprocessing guide to standardize images
3. Create train/validation/test splits as described

# Navigate to project directory
cd dental_ai_project

# Run the training script
python monai_training_script.py

# Results will be saved in dental_ai_dataset/results/

The training script includes several configurable parameters:

• Model architecture (DenseNet121 or ResNet50)
• Learning rate
• Batch size
• Number of epochs
• Data augmentation options

Modify these in the Config class at the beginning of the script.

If you encounter issues during setup or training:

1. Refer to the troubleshooting_guide.md for common problems and solutions
2. Check that your dataset follows the expected directory structure
3. Verify that all dependencies are correctly installed

• MONAI Documentation: https://docs.monai.io/
• PyTorch Tutorials: https://pytorch.org/tutorials/
• Dental X-ray Datasets: See dataset_research.md for options

When using these materials or the trained models in research or clinical applications, please cite:

MONAI Consortium. (2020). MONAI: Medical Open Network for AI. https://monai.io/

For specific dental X-ray datasets, refer to the citation requirements in dataset_research.md.

For questions or assistance with these materials, please contact the instructor, Francisco Barbosa, email: cisco@periospot.com.

Prepared by Manus AI for dental AI training class, June 2025.