Module 1: Course preparation#

Setting up tools and dependencies#

Perform the instructions in the following sections to prepare your machine for the course. You should then have a Python environment setup for GPU computing, Deconvolution and Deep Learning with the following tools

  • Mambaforge: An efficient package and environment manager for scientific computing.

  • Napari and devbio-napari: A 5D image viewer plus a bundle of plugins for processing microscopy images.

  • OpenCl, CLIJ, and Clesperanto for GPU computing with OpenCl

  • CuPy for GPU computing with Cuda (note Mac m1 does not support Cuda)

  • TNIA-Python for projections and helper functions

  • Clij2-fft for non-circulant Richardson Lucy Deconvolution with Total Variation Regularization

  • Tensorflow, Stardist and CSBDeep for deep learning segmentation and restoration

  • Raster-Geometry to create simple simulated objects

Getting the examples#

The examples for this course are organized into a jupyter-book. To get the most out of the course

  1. Open the online book in a web browser and study the example notebooks. The online version of the course can be found here. The code examples are a rendering of code that has been run previously. This view shows the expected outputs of the examples. (We assume if you are reading this you have already found the online version of the course).

  2. Also download the code to your personal device or a cloud VM and make sure you can run it successfully. You can get the code on github at True-North-Intelligent-Algorithms/deconvolution-gpu-dl-course.

You can use the green ‘code’ button to either get the url of the code (to clone using git) or download the code as a .zip.

github

To run the code locally browse into the folder where the code resides. IE if the code is on your desktop

cd Desktop/deconvolution-gpu-dl-course

Then start your preferred jupyter notebook development tool.

For example if you use jupyter lab simply type

jupyter lab

Or if you use Visual Studio Code

code .