AI in Medical Imaging (X-ray, MRI) Full Syllabus

MODULE 1: Introduction to Medical Imaging & AI

1. What is Medical Imaging?
   • Types: X-ray, MRI, CT, Ultrasound, PET
   • Why imaging is important in healthcare?
2. Basics of Artificial Intelligence in Healthcare
   • Difference between AI, ML & Deep Learning
   • Use-cases of AI in hospitals and diagnostics
3. Role of AI in Medical Imaging
   • Detection, segmentation, classification
   • Faster diagnosis and improved accuracy

MODULE 2: Medical Imaging Data Fundamentals

1. Understanding DICOM format (Digital Imaging and Communications in Medicine)
   • How hospitals store images
   • How to read DICOM images in Python
2. Image Types and Preprocessing
   • Grayscale images
   • Image normalization, resizing
   • Removing noise (denoising)
3. Annotation & Labeling
   • What is annotation in medical imaging?
   • Tools: Labelbox, CVAT, MedSeg
   • Types of annotations: Bounding box, Mask, Polygon

MODULE 3: Deep Learning Basics for Medical Imaging

1. Introduction to CNN (Convolutional Neural Networks)
   • Filters, strides, pooling explained in simple terms
   • Why CNNs work so well for images
2. Image Classification
   • Use CNNs to detect pneumonia in X-ray
   • Binary vs Multi-class classification
3. Image Segmentation
   • Introduction to U-Net architecture
   • Use-case: Tumor segmentation in MRI
4. Data Augmentation Techniques
   • Flip, rotate, zoom, crop — and why it’s important

MODULE 4: Building AI Models for X-ray Analysis

1. Chest X-ray Case Study
   • Dataset: NIH ChestX-ray14 or RSNA Pneumonia Detection
   • Preprocessing + Augmentation
   • Model building with TensorFlow or PyTorch
2. Detecting Diseases from X-ray
   • Pneumonia
   • Tuberculosis
   • COVID-19
3. Model Evaluation Metrics
   • Accuracy, Precision, Recall, F1-Score
   • AUC-ROC Curve
   • Confusion Matrix (explained with examples)

MODULE 5: Building AI Models for MRI Analysis

1. Brain Tumor Detection from MRI
   • Dataset: BRATS (Brain Tumor Segmentation)
   • Multi-modal MRI images: T1, T2, FLAIR
2. 3D Imaging Concepts
   • Slicing 3D MRI data for 2D CNNs
   • Working with volumetric data
3. Segmentation Using U-Net / 3D U-Net
   • Mask generation
   • Post-processing and visual results
4. Alzheimer’s Detection Using MRI
   • Case study using CNN + metadata
   • Explainability with Grad-CAM

MODULE 6: Tools, Libraries & Platforms

1. Tools & Libraries
   • Python, TensorFlow, Keras, PyTorch
   • OpenCV, Scikit-Image, SimpleITK
   • MONAI (Medical Open Network for AI)
2. Annotation & Viewer Tools
   • ITK-SNAP, 3D Slicer
   • DICOM viewers like OsiriX, RadiAnt
3. Deployment Basics
   • Exporting models to ONNX
   • Creating a web dashboard using Streamlit

MODULE 7: Challenges & Ethical Concerns

1. Data Privacy (HIPAA, GDPR)
2. Explainability in AI Diagnosis
3. Handling Imbalanced Datasets
4. Bias in medical AI models
5. FDA approval & real-world deployment