torchio

Medical imaging processing for AI applications.

What is torchio?

Medical imaging processing for AI applications.

Overview of torchio

What it is

TorchIO is a Python package for reading, preprocessing, sampling, augmenting, and writing 3D medical images in deep learning applications using PyTorch. It includes intensity and spatial transforms for data augmentation and preprocessing, such as random affine transformations and domain-specific simulations of MRI artifacts like bias field inhomogeneity and k-space motion artifacts. The package was inspired by NiftyNet.

Key features

  • Efficient reading, preprocessing, sampling, augmentation, and writing of 3D medical images - Intensity transforms: random blur, noise, bias field artifact, motion artifact, spike artifact, ghosting artifact - Spatial transforms: random flip, affine transformation, elastic transformation - Patch-based sampling via a queue for training - Designed for PyTorch-based deep learning workflows

Use cases

  • Preprocessing and augmenting medical imaging datasets for AI model training - Simulating realistic MRI artifacts to improve model robustness - Patch-based training of segmentation or classification models on large 3D volumes

Getting started

See the repository README for installation instructions and a Hello, World! example.

License and pricing

Licensed under Apache-2.0. The package is open source and free to use.

Based on the project's public README and repository metadata.

Add this badge to your README

Featured on YunoDB Badge styles & help

Maintain this project? Show that it is featured on YunoDB.

Markdown[![Featured on YunoDB](https://yunodb.com/badge/torchio-projecttorchio.svg)](https://yunodb.com/tools/torchio-projecttorchio?utm_source=github&utm_medium=badge)
HTML<a href="https://yunodb.com/tools/torchio-projecttorchio?utm_source=github&utm_medium=badge"><img src="https://yunodb.com/badge/torchio-projecttorchio.svg" alt="Featured on YunoDB" height="20"></a>