Introduction to Deep Learning for Microscopy

Do you want to learn how to use deep learning for solving your microscopy image analysis tasks? Then you have come to the right spot!

We offer materials for an introductory course on the topic, containing video lectures, exercises that show step by step how to build deep learning models for microscopy with PyTorch and advanced examples that explain how to use several of the most popular deep learning based tools.

Sparked your interest?

• Check out the content!
• See how you can work on the exercises on publicly available resources or how to set them up on your own computer.
• Check out the recommended prior knowledge and further reading.

Content

This course consists of lectures and exercises that teach the background of deep learning for image analysis and show applications to classification and segmentation analysis problems.

Video Lectures

Coming soon.

PyTorch Exercises

These exercises demonstrate how to use PyTorch for image analysis for several tasks on a microscopy image analysis problem. We recommend that you work on these exercises in this order:

1. data-visualization-napari: will give you an overview of the data used in the exercise and give an introduction to napari, which is the image viewer we will use throughout the exercises.
2. cell_classification/pytorch/train_infection_classifier: will teach you how to train and apply a neural network for cell classification.
3. nucleus_segmentation/bioimageio/pretrained_segmentation: will teach you how to apply a pretrained segmentation network from bioimageio.io.
4. cell_segmentation/pytorch/train_cell_segmentation: will teach you how to train and apply a U-Net for cell segmentation.

We provide two additional exercises that show you how to annotate your own data, so that you can create training data for applying what you have learned to your own analysis problems:

• data_annotation/segment_anything: will teach you how to use Segment Anything for Microscopy to interactively annotate training data for segmentation.
• data_annotation/class_annotation: will teach you how to annotate data for classifcation with napari.

Exercises for Other Tools

Coming soon.

Getting Started

On BAND

BAND is a online service for image analysis. It is free of charge and it offers a pre-installed environment for the exercises. Follow this link for a step-by-step guide for how to run the course on BAND.

On your own Computer

Coming soon.

On Kaggle

Coming soon.

Recommended Knowledge and Further Reading

Coming soon.