The open-source Python library AUCMEDI is an intuitive, high-level API allowing fast setup of medical image classification pipelines with state-of-the-art methods in just a few lines of code.

Deep Learning

Library of deep neural networks for binary, multi-class as well as multi-label classification ranging from classics like ResNet to modern Transformer architectures.

Class Imbalance

Efficient methods against small dataset biases as well as class imbalances using class weights and modern loss functions such as focal loss.

Build for Medical Data

Incorporation of clinical metadata and extensive data augmentation for biomedical images based on Albumentations and batchgenerators.

Image Preprocessing

Various preprocessing methods for complexity reduction, such as color conversions, windowing, filtering, resizing and normalization.

Ensemble Learning

Simple integration of Ensemble Learning techniques like test-time augmentation, bagging via cross-validation or stacking via logistic regressions.

Explainable AI

Opening the blackbox! Transparency for opaque decision-making processes using activation maps such as Grad-CAMā  or LIME.


Automated Machine Learning to ensure easy deployment, integration and maintenance of complex medical image classification pipelines.


Documentation, examples, unittesting, continuous integration, continuous development, open-source, active support, and more buzzwords.


Combining standardization of the best methods in the literature with open interfaces for 3rd party contributions.