11.2. How to parallelize loops#

In image processing, we frequently apply the same algorithm on a large batch of images. In this paragraph, we propose to use joblib to parallelize loops. Here is an example of such repetitive tasks:

import skimage as ski

def task(image):
    Apply some functions and return an image.
    image = ski.restoration.denoise_tv_chambolle(
        image[0][0], weight=0.1, channel_axis=-1
    fd, hog_image = ski.feature.hog(
        pixels_per_cell=(16, 16),
        cells_per_block=(1, 1),
    return hog_image

# Prepare images
hubble = ski.data.hubble_deep_field()
width = 10
pics = ski.util.view_as_windows(
    hubble, (width, hubble.shape[1], hubble.shape[2]), step=width

To call the function task on each element of the list pics, it is usual to write a for loop. To measure the execution time of this loop, you can use ipython and measure the execution time with %timeit.

def classic_loop():
    for image in pics:

%timeit classic_loop()

Another equivalent way to code this loop is to use a comprehension list which has the same efficiency.

def comprehension_loop():
    [task(image) for image in pics]

%timeit comprehension_loop()

joblib is a library providing an easy way to parallelize for loops once we have a comprehension list. The number of jobs can be specified.

from joblib import Parallel, delayed
def joblib_loop():
    Parallel(n_jobs=4)(delayed(task)(i) for i in pics)

%timeit joblib_loop()