Slice and Merge provides a tool for slicing and merging large images. This can make it easier to process large images. The main class is TiledImage. It represents a 3D array as a 5D array of tiles. Each tile can be accessed by its index. Also a list of all tiles is accessible. If each tile has been processed separately, the merge method can easily combine these tiles into a full-size image.

Example of use:

from PIL import Image
from slice_merge.slice_merge import TiledImage
path_to_image = '<path to your image>'
image = np.array(Image.open(path_to_image))
sliced_image = TiledImage(image)

Now, a list of tiles can be retrieved by

sliced_image.list_tiles()

Let's apply a simple function that returns the Red channel of a RGB input image. First, define the function:

def get_red_channel(img):
    return img[:, :, 0]

Then apply the function to every tile:

red_tiles = [get_red_channel(tile) for tile in sliced_image.list_tiles()]

or

red_tiles = sliced_image.apply(get_red_channel, parallel=False)

And finally combine results into the full-size image:

red_full_size = sliced_image.merge(red_tiles)