Image analysis and visualisation with Python and napari

Organised by Adam Tyson.

Overview

Many areas of cancer research now rely on the acquisition and analysis of complex, multidimensional images. Existing software may not have all the necessary features or may not be easy to fully automate.

The extensive ecosystem of packages for scientific computing, image analysis and machine learning (numpy, scipy, scikit-image, scikit-learn, tensorflow, PyTorch etc.) greatly simplify the analysis of complex images in Python. The development of napari, the multidimensional image viewer now makes it easy to combine image analysis and visualisation.

This one-day course will teach you how to develop your own, custom image analysis software, beginning with traditional computer vision approaches. We will also introduce some more advanced machine-learning based tools that can be included in your workflows. Throughout the course we will be using napari to visualise and explore the data. At the end we will cover how to turn your image analysis scripts into easy-to-use napari plugins.

To make the most of this course, you should have some experience with Python and basic image analysis (e.g. using other graphical software). You will need a laptop with conda (miniconda or anaconda) installed. If you need any help prior to the course (e.g. if you have a different Python installation) contact Adam Tyson (adam.tyson01@icr.ac.uk).

Contents

• Loading image file formats into Python
• Image visualisation using napari
• Basic image analysis using numpy, scipy and scikit-image
• Advanced machine-learning based image analysis
• Developing user-friendly napari image analysis plugins