MedImages.jlMedical Image Processing in Julia

A comprehensive Julia package for loading, transforming, and analyzing 3D/4D medical images with proper spatial metadata handling

Get Started

API Reference

View on GitHub

Complete I/O Support

Load and save NIfTI files with full spatial metadata preservation. HDF5 storage for efficient dataset management.

Spatial Transformations

Rotation, cropping, padding, scaling, and translation with automatic origin/spacing adjustment.

Resampling & Registration

Resample to new spacing or register images to a common coordinate space with multiple interpolation methods.

Orientation Management

Convert between 8 standard orientations (LPS, RAS, etc.) with proper coordinate system handling.

What is MedImages.jl?

MedImages.jl is a Julia package for standardizing the handling of 3D and 4D medical images. It provides a unified data structure that combines voxel data with comprehensive spatial and clinical metadata, ensuring that transformations maintain physical accuracy.

Key Features

Unified Data Structure: The MedImage struct encapsulates voxel data, spatial metadata (origin, spacing, direction), and clinical metadata in a single container.
Coordinate System Aware: All operations maintain proper relationships between array indices and physical coordinates.
Multiple Interpolation Methods: Nearest neighbor, linear, and B-spline interpolation for different use cases.
Orientation Support: Work with 8 standard orientations and convert between them seamlessly.
ITK/SimpleITK Compatible: Uses ITKIOWrapper for I/O, ensuring compatibility with the broader medical imaging ecosystem.

Quick Start

Installation

julia

using Pkg
Pkg.add("MedImages")

Basic Usage

julia

using MedImages

# Load a CT scan
ct = load_image("/path/to/scan.nii.gz", "CT")

# Inspect properties
println("Dimensions: ", size(ct.voxel_data))
println("Spacing (mm): ", ct.spacing)
println("Origin (mm): ", ct.origin)

# Resample to isotropic spacing
isotropic = resample_to_spacing(ct, (1.0, 1.0, 1.0), Linear_en)

# Save result
create_nii_from_medimage(isotropic, "/output/isotropic_ct")

Supported Modalities

Modality	Type	Subtypes
CT	`CT_type`	CT_subtype
MRI	`MRI_type`	T1, T2, FLAIR, ADC, DWI
PET	`PET_type`	FDG, PSMA

Available Transformations

Function	Description
`crop_mi`	Extract a region of interest
`pad_mi`	Add voxels around the image
`rotate_mi`	Rotate around an axis
`translate_mi`	Shift the image origin
`scale_mi`	Resize the image
`resample_to_spacing`	Change voxel spacing
`resample_to_image`	Register to another image
`change_orientation`	Convert between orientations

Integration

MedImages.jl is designed to work with:

ITKIOWrapper.jl - For DICOM and NIfTI I/O
MedEye3d.jl - For 3D visualization
MedPipe3D.jl - For processing pipelines

Documentation Sections

Getting Started - Installation and setup instructions
Tutorials - Step-by-step guides for common workflows
Code Examples - Ready-to-use code snippets
API Reference - Complete function documentation
Data Structures - Type and enum documentation

Contributing

Contributions are welcome! Please visit our GitHub repository to:

Report issues
Submit pull requests
Suggest new features

License

MedImages.jl is released under the MIT License.

Part of the JuliaHealth organization