pyclesperanto-prototype 0.4.0

OpenCL based GPU-accelerated image processing (an early prototype)

pyclesperanto

pyclesperanto is a prototype for clEsperanto - a multi-platform multi-language framework for GPU-accelerated image processing. It uses OpenCL kernels from CLIJ

Right now, this is very preliminary.

Installation

• Get a python environment, e.g. via mini-conda
• Install pyopencl.

If installation of pyopencl for Windows fails, consider downloading a precompiled wheel (e.g. from here ) and installing it manually:

pip install pyopencl-2019.1.1+cl12-cp37-cp37m-win_amd64.whl

Afterwards, install pyclesperanto:

pip install pyclesperanto-prototype

Troubleshooting installation

If you receive an error like

DLL load failed: The specified procedure could not be found.

Try downloading and installing a pyopencl with a lower cl version, e.g. cl12 : pyopencl-2020.1+cl12-cp37-cp37m-win_amd64

Example code

A basic image procressing workflow loads blobs.gif and counts the number of gold particles:

import pyclesperanto_prototype as cle

from skimage.io import imread, imsave

# initialize GPU
cle.select_device("GTX")
print("Used GPU: " + cle.get_device().name)

# load data
image = imread('https://imagej.nih.gov/ij/images/blobs.gif')
print("Loaded image size: " + str(image.shape))

# push image to GPU memory
input = cle.push(image)
print("Image size in GPU: " + str(input.shape))

# process the image
inverted = cle.subtract_image_from_scalar(image, scalar=255)
blurred = cle.gaussian_blur(inverted, sigma_x=1, sigma_y=1)
binary = cle.threshold_otsu(blurred)
labeled = cle.connected_components_labeling_box(binary)

# The maxmium intensity in a label image corresponds to the number of objects
num_labels = cle.maximum_of_all_pixels(labeled)

# print out result
print("Num objects in the image: " + str(num_labels))

# for debugging: print out image
print(labeled)

# for debugging: save image to disc
imsave("result.tif", cle.pull(labeled))

More examples can be found in the /demo/ directory.