Created
March 11, 2016 19:34
-
-
Save nlintz/f5bde5df8eee57472954 to your computer and use it in GitHub Desktop.
Utilities for converting an image into a sequence of non-overlapping patches and back to an image again
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
import numpy as np | |
from skimage.util import view_as_blocks | |
def round_down(num, divisor): | |
return num - (num % divisor) | |
def crop_center(img, new_rows, new_cols): | |
rows, cols, c = img.shape | |
top = (rows - new_rows)/2 | |
bot = (rows + new_rows)/2 | |
left = (cols - new_cols)/2 | |
right = (cols + new_cols)/2 | |
return img[top:bot, left:right, :] | |
def crop_image_to_patch(img, patch_shape): | |
""" center crops an image so its shape can be broken down | |
into patch_shaped chunks | |
img: [rows, cols, nchannels] | |
patch_shape: (rows, cols, nchannels) | |
""" | |
rows, cols, c = img.shape | |
new_rows = round_down(rows, patch_shape[0]) | |
new_cols = round_down(cols, patch_shape[1]) | |
return crop_center(img, new_rows, new_cols) | |
def image_to_patches(img, patch_shape): | |
""" | |
converts an image into M patch_shaped chunks | |
arguments -- | |
img: [rows, cols, nchannels] | |
patch_shape: (rows, cols, nchannels) | |
returns -- | |
[M, patch_shape[0], patch_shape[1], nchannels] | |
""" | |
blocks = view_as_blocks(img, patch_shape) | |
return blocks.reshape((-1,) + patch_shape, order="F") | |
def patches_to_image(patches, patch_shape, img_shape): | |
""" | |
converts result from image_to_patches back into an image of the | |
original image's shape | |
patches: [n_patches, rows, cols, nchannels] | |
patch_shape: (rows, cols, nchannels) | |
img_shape: (rows, cols, nchannels) | |
""" | |
patches_per_col = img_shape[0] / patch_shape[0] | |
res = [] | |
for i in range(0, patches_per_col): | |
res.append(np.concatenate(patches[i::patches_per_col], axis=1)) | |
return img | |
if __name__ == "__main__": | |
# Example Script | |
import matplotlib.pyplot as plt | |
patch_size = (2, 2, 3) | |
img = np.repeat(np.repeat(np.random.rand(4, 4, 3), 2, axis=0), 2, axis=1) | |
img = crop_image_to_patch(img, patch_size) | |
patches = image_to_patches(img, patch_size) | |
plt.subplot(311) | |
plt.imshow(img, interpolation="None") | |
plt.subplot(312) | |
plt.imshow(np.concatenate(patches), interpolation="None") | |
plt.subplot(313) | |
plt.imshow(patches_to_image(patches, patch_size, img.shape), interpolation="None") | |
plt.show() |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment