There are many filters that are designed to work with gray-scale images but not
with color images. To simplify the process of creating functions that can adapt
to RGB images, scikit-image provides the
To actually use the
adapt_rgb decorator, you have to decide how you want to
adapt the RGB image for use with the gray-scale filter. There are two
Below, we demonstrate the use of
adapt_rgb on a couple of gray-scale
We can use these functions as we would normally use them, but now they work with both gray-scale and color images. Let’s plot the results with a color image:
from skimage import data from skimage.exposure import rescale_intensity import matplotlib.pyplot as plt image = data.astronaut() fig = plt.figure(figsize=(14, 7)) ax_each = fig.add_subplot(121, adjustable='box-forced') ax_hsv = fig.add_subplot(122, sharex=ax_each, sharey=ax_each, adjustable='box-forced') # We use 1 - sobel_each(image) # but this will not work if image is not normalized ax_each.imshow(rescale_intensity(1 - sobel_each(image))) ax_each.set_xticks(), ax_each.set_yticks() ax_each.set_title("Sobel filter computed\n on individual RGB channels") # We use 1 - sobel_hsv(image) but this will not work if image is not normalized ax_hsv.imshow(rescale_intensity(1 - sobel_hsv(image))) ax_hsv.set_xticks(), ax_hsv.set_yticks() ax_hsv.set_title("Sobel filter computed\n on (V)alue converted image (HSV)")
Notice that the result for the value-filtered image preserves the color of the original image, but channel filtered image combines in a more surprising way. In other common cases, smoothing for example, the channel filtered image will produce a better result than the value-filtered image.
You can also create your own handler functions for
adapt_rgb. To do so,
just create a function with the following signature:
def handler(image_filter, image, *args, **kwargs): # Manipulate RGB image here... image = image_filter(image, *args, **kwargs) # Manipulate filtered image here... return image
adapt_rgb handlers are written for filters where the image is
the first argument.
As a very simple example, we can just convert any RGB image to grayscale and then return the filtered result:
It’s important to create a signature that uses
to pass arguments along to the filter so that the decorated function is
allowed to have any number of positional and keyword arguments.
Finally, we can use this handler with
adapt_rgb just as before:
@adapt_rgb(as_gray) def sobel_gray(image): return filters.sobel(image) fig = plt.figure(figsize=(7, 7)) ax = fig.add_subplot(111, sharex=ax_each, sharey=ax_each, adjustable='box-forced') # We use 1 - sobel_gray(image) # but this will not work if image is not normalized ax.imshow(rescale_intensity(1 - sobel_gray(image)), cmap=plt.cm.gray) ax.set_xticks(), ax.set_yticks() ax.set_title("Sobel filter computed\n on the converted grayscale image") plt.show()
A very simple check of the array shape is used for detecting RGB
adapt_rgb is not recommended for functions that support
3D volumes or color images in non-RGB spaces.
Total running time of the script: ( 0 minutes 0.878 seconds)