Image Analysis in Python with SciPy and Scikit-Image

Presented by

Juan Nunez-Iglesias juan.n@unimelb.edu.au
Emmanuelle Gouillart emmanuelle.gouillart@nsup.org, @EGouillart

This tutorial can be found online at https://github.com/scikit-image/skimage-tutorials

Please launch the IPython notebook from the root of the repository.

From telescopes to satellite cameras to electron microscopes, scientists are producing more images than they can manually inspect. This tutorial will introduce automated image analysis using the "images as NumPy arrays" abstraction, run through various fundamental image analysis operations (filters, morphology, segmentation), and then focus on solving real-world problems through interactive demos.

Image analysis is central to a boggling number of scientific endeavors. Google needs it for their self-driving cars and to match satellite imagery and mapping data. Neuroscientists need it to understand the brain. NASA needs it to map asteroids and save the human race. It is, however, a relatively underdeveloped area of scientific computing.

The goal is for attendees to leave the tutorial confident of their ability to extract information from images in Python.

Prerequisites

All of the below packages, including the non-Python ones, can be found in the Anaconda Python distribution, which can be obtained for free. Alternatively, pip install [packagename] should work.

Required packages

scikit-image (0.11 or higher)

Required for scikit-image:

Python (>=2.6 required, 3.4 recommended)
numpy (>=1.6 required, 1.7 recommended)
scipy (>=0.10 required, 0.13 recommended)

Required for image reading and viewing:

matplotlib (>=1.1.0 required, 1.2 recommended)

Example images

scikit-image ships with some example images in skimage.data. For this tutorial, we will additionally make use of images included in the skimage-tutorials/images folder. If you're reading this on your computer, you already have these images downloaded.

Introduction

Gallery, User Guide and API reference
Subpackages, see help(skimage)

The relationship of skimage with the Scientific Python eco-system

numpy (with skimage as the image processing layer)
scipy (in combination with ndimage, a subpackage of scipy)
sklearn (machine learning + feature extraction)
opencv for speed (e.g. in a factory setting)
novice (for teaching)

Schedule

(If the links below don't work, make sure you start the notebook in the root of the repository, not in 2015-euroscipy.)

11:00--11:45 (Juan)

11:45--12:30 (Emmanuelle)

Segmentation

bonus material

Choose your own adventure!

Warping -- Mona Lisa de-warping competition inside!
Analyzing micro arrays
Hands-on exercises -- the best brain teasers from StackOverflow

Even more lectures here and here.