Crosswalk¶
Meet The Beatles at Abbey Road for Image Processing.
Roland Bruggmann, 2018 https://github.com/brugr9/Crosswalk
In [1]:
%matplotlib inline
In [2]:
from skimage.io import imread
import numpy as np
import matplotlib.pyplot as plt
silhouette = imread('images/beatles-abbey-road-silhouette.jpg', True)
cover = imread('images/beatles-abbey-road-cover.jpg', True)
# The Beatles: George Harrison - Paul McCartney - Ringo Starr - John Lennon
george = imread('images/beatles-abbey-road-silhouette-1-george.jpg', True)
paul = imread('images/beatles-abbey-road-silhouette-2-paul.jpg', True)
ringo = imread('images/beatles-abbey-road-silhouette-3-ringo.jpg', True)
john = imread('images/beatles-abbey-road-silhouette-4-john.jpg', True)
thebeatles = [[george, 'coral'], [paul, 'deepskyblue'], [ringo, 'fuchsia'], [john, 'gold']]
1. Template Matching¶
In [3]:
from skimage.feature import match_template
def matchandplot( image, template, color ):
"Match a template using [...] normalized correlation. \
The output is an array with values between -1.0 and 1.0. \
The value at a given position corresponds to the correlation coefficient between the image and the template. \
[...] Matches correspond [...] to the top-left corner of the template.\
To find the best match one must search for peaks in the response (output) image."
response = match_template(image, template)
# search for peaks in the response.
ij = np.unravel_index(np.argmax(response), response.shape)
x, y = ij[::-1]
# plot
fig = plt.figure(figsize=(12, 2))
ax1 = plt.subplot(1, 2, 1)
ax2 = plt.subplot(1, 2, 2, adjustable='box-forced')
ax1.imshow(template, cmap=plt.cm.gray)
ax1.set_axis_off()
ax1.set_title('Template')
ax2.imshow(image, cmap=plt.cm.gray)
ax2.set_axis_off()
ax2.set_title('Image')
# highlight matched region
h, w = template.shape
rect = plt.Rectangle((x, y), w, h, edgecolor=color, facecolor='none', lw=2)
ax2.add_patch(rect)
plt.show()
return
for member in thebeatles:
matchandplot( silhouette, member[0], member[1] )
In [ ]: