Image Analysis in Python#
This half-day course will introduce the use of Python packages to analyse and visualise image data common to neuroscience.
Interactive workflows in Python:
IPython for efficient data exploration
Basic image analysis techniques:
Utilizing NumPy, and scikit-image libraries
Visualization and exploration with napari
Processing large datasets effectively:
Introduction to Dask for parallel computing
Demonstration using BrainSaw data
Tools for processing histology data:
BrainGlobe tools for image registration and segmentation
If time allows, we will also look at using convolutional neural networks for tricky segmentation problems.
In advance of the course#
Before attending the course, please install conda if you don’t already have it installed, and then run the following to install all relevant packages:
conda create --name image-analysis-python python=3.10 nb_conda_kernels -y conda activate image-analysis-python git clone https://github.com/neuroinformatics-unit/image-analysis-python cd image-analysis-python pip install -r requirements.txt
To speed things up on the day, you may wish to download the data in advance. To do this:
Start jupyter lab (
Open up the first notebook (e.g.
Set the conda environment (should be
image-analysis-pythonbased on the above commands)
Run the first code cell (the one that says Run the following cell to download the data in advance above it!)
Repeat for the other notebooks
dask_cellfinder notebook has the largest sample data, so this is probably the best one to download in advance.
Even on a fast (e.g. UCL) network, it may take ~1hr to download.