Skip to main content
Ctrl
+
K
PoL Bio-Image Analysis Training School on GPU-Accelerated Image Analysis
Monday
Course preparation
Clesperanto
List and select devices
GPU arrays
Apply operations on data
Image Filtering and Segmentation
Object measurements and quantifications
Custom kernel execution
Benchmarking
Using clesperanto from the Napari Assistant
The Napari Assistant
Generating Jupyter Notebooks from the Napari Assistant
cupy
Basics of Cupy
Cupy as drop-in replacement for numpy
Image filtering using cupy
Custom kernels
napari integration
Benchmarking affine transforms using numpy, cupy and clesperanto.
Intro to Deconvolution and Restoration
Setup environment
Implementing the forward model (Convolution) with cupy
Nuclei Deconvolution and Compare intensities to ground truth
Nuclei Deconvolution and Segmentation
Edge handling
Edge handling experiments
Deconvolve microtubules phantom
A Notebook showing ‘reverse Deconvolution’ a.k.a. PSF Distilling
Dask deconvolution
Running deconvolution on the ZIH cluster
Tuesday
High-performance-computing
Using Clesperanto on Taurus
Executing clesperanto on the TU Dresden HPC
Loading ‘blobs’
Exercises: GPU-accelerated image processing on HPC
Introduction to Pytorch
Versions
Data exploration
Creation of a dataset
Processing batches of data
Training a Unet
Going for the GPU
Training with logging
Training with checkpoints
Pytorch lightning
Wednesday
AI segmentation and denoising
Training a 2D Unet model
(Probabilistic) Noise2Void (2D)
Noise2Void (3D)
Lazy and parallel bio-image processing using DASK
Practical 1: Dask basics
Practical 2: Dask with images
Practical 3: Virtual stack visualization and explorative analysis
Repository
Open issue
Search
Error
Please activate JavaScript to enable the search functionality.
Ctrl
+
K
