Examples

This section provides practical examples demonstrating how to use MedTDA for various tasks.

Overview 

The examples are organized by complexity and use case:

Basic Usage - Get started with simple feature extraction
Advanced Preprocessing - Custom preprocessing workflows
Batch Workflows - Process multiple images efficiently
CLI Workflows - Command-line usage patterns
Tutorial Notebook - Interactive Jupyter notebook

All examples use realistic medical imaging scenarios and follow best practices.

Quick Links 

Example	Description
Basic Usage	Simple feature extraction, masks, different formats
Advanced Preprocessing	Custom pipelines, ROI analysis, multi-step workflows
Batch Workflow	CSV batch processing, parallel execution, result analysis
CLI Workflows	CLI patterns, scripting, pipeline integration
Interactive Tutorial	Interactive Jupyter notebook tutorial

Example Gallery 

Getting Started 

If you’re new to MedTDA, start with:

Basic Usage - Learn the fundamentals
Quick Start - 5-minute quick start guide
Advanced Preprocessing - More control over preprocessing

For Production Use 

When deploying MedTDA:

Batch Workflow - Efficient batch processing
CLI Workflows - Integrate with existing pipelines

Example Data 

Most examples use placeholder paths. To run them:

Option 1: Use your own data

Replace paths with your medical images:

# Replace this
image_path = 'data/patient001.nii.gz'

# With your actual path
image_path = '/path/to/your/scan.nii.gz'

Option 2: Generate synthetic data

For testing purposes:

import numpy as np
from medtda.loaders import save_image

# Create synthetic 3D image
synthetic_image = np.random.rand(64, 64, 64)
save_image(synthetic_image, 'test_image.nii.gz')

# Create synthetic mask
synthetic_mask = (np.random.rand(64, 64, 64) > 0.5).astype(np.uint8)
save_image(synthetic_mask, 'test_mask.nii.gz')

Option 3: Download sample data

See Installation for links to public medical imaging datasets.

Example Categories 

By Task 

Classification:

Basic Usage - Extract features for ML
Batch Workflow - Process training/test sets

Regression:

Basic Usage - Continuous outcome prediction
Advanced Preprocessing - Optimize preprocessing

Exploratory Analysis:

Basic Usage - Quick feature inspection

Large-Scale Processing:

Batch Workflow - Hundreds/thousands of images
CLI Workflows - Automation and scripting

By Image Type 

CT Scans:

Advanced Preprocessing - CT windowing examples
Basic Usage - CT-specific workflows

MRI:

Basic Usage - Multi-sequence MRI
Advanced Preprocessing - Normalization strategies

Microscopy:

Basic Usage - 2D image analysis

Multi-Modal:

Batch Workflow - Process different modalities
Advanced Preprocessing - Modality-specific preprocessing

Code Patterns 

Common Import Pattern 

Most examples start with:

from medtda import FeatureExtractor, Preprocessor, BarcodeExtractor
from medtda.loaders import load_image, load_mask
from medtda.vectorizers import persistence_image
import numpy as np

Basic Workflow 

Standard processing pattern:

# 1. Load data
image, metadata = load_image('scan.nii.gz')
mask, _ = load_mask('mask.nii.gz')

# 2. Configure extractor
extractor = FeatureExtractor(
    normalize=True,
    vectorization_method='PersImage'
)

# 3. Extract features
features = extractor.execute(image, mask)

# 4. Use features
# ... machine learning, analysis, etc.

Error Handling 

Robust processing:

from medtda import FeatureExtractor

try:
    extractor = FeatureExtractor()
    features = extractor.execute(image_path, mask_path)
except FileNotFoundError as e:
    print(f"File not found: {e}")
except ValueError as e:
    print(f"Invalid data: {e}")
except Exception as e:
    print(f"Unexpected error: {e}")

Next Steps 

After exploring examples:

User Guide - Comprehensive user guide
API Reference - Detailed API reference
TDA Theory - Understand the theory
Frequently Asked Questions - Common questions

Contributing Examples 

Have a useful example? Contribute!

See Contributing to MedTDA for guidelines on:

Submitting new examples
Improving existing examples
Adding tutorial notebooks
Sharing use cases

Tips for Using Examples 

Start Simple

Begin with Basic Usage before complex workflows
Adapt to Your Data

Modify paths, parameters, and methods for your specific case
Experiment

Try different preprocessing and vectorization options
Check Performance

Monitor computation time for large datasets
Validate Results

Always verify that features make sense for your application