Import dask image is with an underscore, like this example:
import dask_image.imread import dask_image.ndfilters
We highly recommend checking out the dask-image-quickstart.ipynb notebook
(and any other dask-image example notebooks) at the dask-examples repository.
You can find the dask-image quickstart notebook in the
of this repository:
The direct link to the notebook file is here:
All the example notebooks are available to launch with mybinder and test out interactively.
An Even Quicker Start¶
You can read files stored on disk into a dask array
by passing the filename, or regex matching multiple filenames
filename_pattern = 'path/to/image-*.png' images = dask_image.imread.imread(filename_pattern)
If your images are parts of a much larger image, dask can stack, concatenate or block chunks together: http://docs.dask.org/en/latest/array-stack.html
Calling dask-image functions is also easy.
import dask_image.ndfilters blurred_image = dask_image.ndfilters.gaussian_filter(images, sigma=10)
Many other functions can be applied to dask arrays. See the dask_array_documentation for more detail on general array operations.
result = function_name(images)
Good places to start include:
The dask-image API documentation: http://image.dask.org/en/latest/api.html
The documentation on working with dask arrays: http://docs.dask.org/en/latest/array.html
Talks and Slides¶
Here are some talks and slides that you can watch to learn dask-image:
2020, Genevieve Buckley’s talk at PyConAU and SciPy Japan
Scipy Japanのトークを見る(プレゼンテーション:英語, 字幕:日本語) Watch the talk at SciPy Japan (presentation in English, captions in Japanese)
2019, John Kirkham’s SciPy talk