[np.ndarray, np.bool]: [A binary image]
"""
# START YOUR CODE HERE ### (You can change anything inside this block)
binary_im = np.zeros_like(im, dtype=np.bool)
### END YOUR CODE HERE ###
return binary_im
if __name__ == "__main__":
# NO NEED TO EDIT THE CODE BELOW.
verbose = True
plt.figure(figsize=(4, 12))
plt.tight_layout()
images_to_visualize = []
for i, impath in enumerate(impaths):
im = utils.read_im(str(impath))
im_binary = create_binary_image(im)
assert im_binary.dtype == np.bool, f"Expected the image to be of dtype np.bool, got {im_binary.dtype}"
angles, distances = utils.find_angle(im_binary)
angle = 0
if len(angles) > 0:
angle = angles[0] * 180 / np.pi
print(f"Found angle: {angle:.2f}")
hough_im = utils.create_hough_line_image(im, angles, distances)
rotated = skimage.transform.rotate(im, angle, cval=im.max())
images_to_visualize.extend([im, im_binary, hough_im, rotated])
image = utils.np_make_image_grid(images_to_visualize, nrow=len(impaths))
utils.save_im("task4d.png", image)
plt.imshow(image, cmap="gray")
import matplotlib.pyplot as plt
import pathlib
import numpy as np
import time
from utils import read_im, save_im, normalize
output_dir = pathlib.Path("image_solutions")
output_dir.mkdir(exist_ok=True)
im = read_im(pathlib.Path("images", "lake.jpg"))
plt.imshow(im)
def convolve_im(im, kernel):
""" A function that convolves im with kernel
Args:
im ([type]): [np.array of shape [H, W, 3]]
kernel ([type]): [np.array of shape [K, K]]
Returns:
[type]: [np.array of shape [H, W, 3]. should be same as im]
"""
assert len(im.shape) == 3
# Rotate kernel
cKernel = kernel.copy()
kLen = len(kernel)
for y in range(kLen):
for x in range(kLen):
cKernel[y][x] = kernel[kLen - y - 1][kLen - x - 1]
import os
import numpy as np
import utils
import matplotlib.pyplot as plt
def magnitude(fft_im):
real = fft_im.real
imag = fft_im.imag
return np.sqrt(real**2 + imag**2)
if __name__ == "__main__":
# DO NOT CHANGE
impath = os.path.join("images", "noisy_moon.png")
im = utils.read_im(impath)
# START YOUR CODE HERE ### (You can change anything inside this block)
# Since the spikes are on a simple horizontal line in the center around [265:275], we
# can simply craete a kernel which removes theese values. Non-shifted, [265:275] becomes [0:10]
# Note that the middle area should be unaffected.
kernel = np.ones(im.shape)
spike_height = 4 # Lines are barely visible at < 4
image_center_x = im.shape[1] // 2
image_center_y = im.shape[0] // 2
center_width = 28 # Lines are visible for > 28
kernel[image_center_y - spike_height //
2:image_center_y + spike_height // 2] = 0
kernel[image_center_y - spike_height // 2:image_center_y + spike_height //