# Read image
impath = os.path.join("images", "lake.jpg")
im = plt.imread(impath)
# Define the convolutional kernels
h_a = np.ones((3, 3)) / 9
h_b = np.array([
[1, 4, 6, 4, 1],
[4, 16, 24, 16, 4],
[6, 24, 36, 24, 6],
[4, 16, 24, 16, 4],
[1, 4, 6, 4, 1],
]) / 256
# Convolve images
smoothed_im1 = convolve_im(im.copy(), h_a)
smoothed_im2 = convolve_im(im, h_b)
# DO NOT CHANGE
assert isinstance(smoothed_im1, np.ndarray), \
f"Your convolve function has to return a np.array. " +\
f"Was: {type(smoothed_im1)}"
assert smoothed_im1.shape == im.shape, \
f"Expected smoothed im ({smoothed_im1.shape}" + \
f"to have same shape as im ({im.shape})"
assert smoothed_im2.shape == im.shape, \
f"Expected smoothed im ({smoothed_im1.shape}" + \
f"to have same shape as im ({im.shape})"
save_im("convolved_im_h_a.jpg", smoothed_im1)
save_im("convolved_im_h_b.jpg", smoothed_im2)
[-1, -1, -1],
[-1, 8, -1],
[-1, -1, -1],
])
h_d = np.array([
[0, -1, 0],
[-1, 5, -1],
[0, -1, 0],
])
# Convolve images
smoothed_im1 = convolve_im(lake_im, h_a)
smoothed_im2 = convolve_im(lake_im, h_b)
edge_im = convolve_im(lake_im, h_c)
sharpened_im = convolve_im(lake_im, h_d)
save_im("convolved_im_h_a.png", smoothed_im1)
save_im("convolved_im_h_b.png", smoothed_im2)
save_im("edge_detection.png", edge_im)
save_im("sharpened.png", sharpened_im)
# DO NOT CHANGE
assert isinstance(smoothed_im1, np.ndarray), \
"Your convolve function has to return a np.array. " +\
"Was: {type(smoothed_im1)}"
assert smoothed_im1.shape == lake_im.shape, \
"Expected smoothed lake_im ({smoothed_im1.shape}" + \
"to have same shape as lake_im ({lake_im.shape})"
assert smoothed_im2.shape == lake_im.shape, \
"Expected smoothed lake_im ({smoothed_im1.shape}" + \
"to have same shape as lake_im ({lake_im.shape})"