09/06/2022, 15:08 Segmentation.

ipynb - Colaboratory

import numpy as np
import tensorflow as tf
from tensorflow import keras
from tensorflow.keras import layers

from numpy import expand_dims
from keras.preprocessing.image import load_img
from keras.preprocessing.image import img_to_array
from keras.preprocessing.image import ImageDataGenerator

import matplotlib.pyplot as plt
from matplotlib import pyplot
from skimage.color import rgb2gray
from skimage.color import rgb2hsv
import matplotlib.pyplot as plt

# load the image
img = load_img('Healthy_1.jpeg')
# convert to numpy array
data = img_to_array(img)

plt.subplot(1, 2, 1)
 
# Displaying the sample image
plt.imshow(data)
 
# Converting RGB image to Monochrome
gray_image = rgb2gray(data)
plt.subplot(1, 2, 2)
 
# Displaying the sample image - Monochrome
# Format
plt.imshow(gray_image, cmap="gray")

https://colab.research.google.com/drive/1lSBY_UgQ754UwEr2IPsSd9RPBfhTNxzt#printMode=true 1/8
09/06/2022, 15:08 Segmentation.ipynb - Colaboratory

Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).

<matplotlib.image.AxesImage at 0x7f5488e37650>

Saved successfully!

hsv_image = rgb2hsv(data)

plt.subplot(1, 2, 2)

# Displaying the sample image - HSV Format

hsv_data_colorbar = plt.imshow(hsv_image)

# Adjusting colorbar to fit the size of the image

plt.colorbar(hsv_data_colorbar, fraction=0.046, pad=0.04)

Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).

<matplotlib.colorbar.Colorbar at 0x7f548881eb10>

from skimage import filters

from skimage.color import rgb2gray

import matplotlib.pyplot as plt

# Setting plot size to 15, 15

plt.figure(figsize=(15, 15))

https://colab.research.google.com/drive/1lSBY_UgQ754UwEr2IPsSd9RPBfhTNxzt#printMode=true 2/8
09/06/2022, 15:08 Segmentation.ipynb - Colaboratory

# Sample Image of scikit-image package

gray_coffee = rgb2gray(data)

 
Saved successfully!
# Computing Otsu's thresholding value

threshold = filters.threshold_otsu(gray_coffee)

# Computing binarized values using the obtained

# threshold

binarized_coffee = (gray_coffee > threshold)*1

plt.subplot(2,2,1)

plt.title("Threshold: >"+str(threshold))

# Displaying the binarized image

plt.imshow(binarized_coffee, cmap = "gray")

# Computing Ni black's local pixel

# threshold values for every pixel

threshold = filters.threshold_niblack(gray_coffee)

# Computing binarized values using the obtained

# threshold

binarized_coffee = (gray_coffee > threshold)*1

plt.subplot(2,2,2)

plt.title("Niblack Thresholding")

# Displaying the binarized image

plt.imshow(binarized_coffee, cmap = "gray")

# Computing Sauvola's local pixel threshold

# values for every pixel - Not Binarized

threshold = filters.threshold_sauvola(gray_coffee)

plt.subplot(2,2,3)

plt.title("Sauvola Thresholding")

# Displaying the local threshold values

https://colab.research.google.com/drive/1lSBY_UgQ754UwEr2IPsSd9RPBfhTNxzt#printMode=true 3/8
09/06/2022, 15:08 Segmentation.ipynb - Colaboratory

plt.imshow(threshold, cmap = "gray")

# Computing Sauvola's local pixel

# threshold values for every pixel - Binarized

binarized_coffee = (gray_coffee > threshold)*1

Saved successfully!
plt.subplot(2,2,4)

plt.title("Sauvola Thresholding - Converting to 0's and 1's")

# Displaying the binarized image

plt.imshow(binarized_coffee, cmap = "gray")

https://colab.research.google.com/drive/1lSBY_UgQ754UwEr2IPsSd9RPBfhTNxzt#printMode=true 4/8
09/06/2022, 15:08 Segmentation.ipynb - Colaboratory

<matplotlib.image.AxesImage at 0x7f5482016050>

Saved successfully!

import numpy as np

import matplotlib.pyplot as plt

from skimage.color import rgb2gray

from skimage import data

from skimage.filters import gaussian

from skimage.segmentation import active_contour

# Sample Image of scikit-image package

img = load_img('Healthy_1.jpeg')

# convert to numpy array

data = img_to_array(img)

gray_astronaut = rgb2gray(data)

# Applying Gaussian Filter to remove noise

gray_astronaut_noiseless = gaussian(gray_astronaut, 1)

# Localising the circle's center at 220, 110

x1 = 220 + 100*np.cos(np.linspace(0, 2*np.pi, 500))

x2 = 100 + 100*np.sin(np.linspace(0, 2*np.pi, 500))

https://colab.research.google.com/drive/1lSBY_UgQ754UwEr2IPsSd9RPBfhTNxzt#printMode=true 5/8
09/06/2022, 15:08 Segmentation.ipynb - Colaboratory

# Generating a circle based on x1, x2

snake = np.array([x1, x2]).T

# Computing the Active Contour for the given image

astronaut_snake = active_contour(gray_astronaut_noiseless,

Saved successfully!
                                 snake)

fig = plt.figure(figsize=(10, 10))

# Adding subplots to display the markers

ax = fig.add_subplot(111)

# Plotting sample image

ax.imshow(gray_astronaut_noiseless)

# Plotting the face boundary marker

ax.plot(astronaut_snake[:, 0],

        astronaut_snake[:, 1],

        '-b', lw=5)

# Plotting the circle around face

ax.plot(snake[:, 0], snake[:, 1], '--r', lw=5)

[<matplotlib.lines.Line2D at 0x7f547d577ed0>]

import matplotlib.pyplot as plt

from skimage.color import rgb2gray

from skimage import data, img_as_float

from skimage.segmentation import chan_vese

fig, axes = plt.subplots(1, 3, figsize=(10, 10))

# Sample Image of scikit-image package

# Sample Image of scikit-image package

https://colab.research.google.com/drive/1lSBY_UgQ754UwEr2IPsSd9RPBfhTNxzt#printMode=true 6/8
09/06/2022, 15:08 Segmentation.ipynb - Colaboratory

img = load_img('Healthy_1.jpeg')

# convert to numpy array

data = img_to_array(img)

gray_astronaut = rgb2gray(data)

 
Saved successfully!
# Computing the Chan VESE segmentation technique

chanvese_gray_astronaut = chan_vese(gray_astronaut,

                                    max_iter=100,

                                    extended_output=True)

ax = axes.flatten()

# Plotting the original image

ax[0].imshow(gray_astronaut, cmap="gray")

ax[0].set_title("Original Image")

# Plotting the segmented - 100 iterations image

ax[1].imshow(chanvese_gray_astronaut[0], cmap="gray")

title = "Chan-Vese segmentation - {} iterations".format(len(chanvese_gray_astronaut[2]))

ax[1].set_title(title)

# Plotting the final level set

ax[2].imshow(chanvese_gray_astronaut[1], cmap="gray")

ax[2].set_title("Final Level Set")

plt.show()

https://colab.research.google.com/drive/1lSBY_UgQ754UwEr2IPsSd9RPBfhTNxzt#printMode=true 7/8
09/06/2022, 15:08 Segmentation.ipynb - Colaboratory

Saved successfully!

https://colab.research.google.com/drive/1lSBY_UgQ754UwEr2IPsSd9RPBfhTNxzt#printMode=true 8/8