Convert Image into Sketch

In Python, an image is just a two-dimensional array of integers. So one can do a couple of matrix manipulations using various python modules in order to get some very interesting effects. In order to convert the normal image to a sketch, we will change its original RGB values and assign its RGB values similar to grey, in this way a sketch of the input image will be generated. 

Approach 1:

1. Import all required modules (numpy, imageio, scipy.ndimage, OpenCV)
2. Take Image input
3. Check RGB value of image and convert into according to RGB values
4. Show finale image output using cv2.imwrite()

Python3

# Python program to Convert Image into sketch


# import all the required modules
import numpy as np
import imageio
import scipy.ndimage
import cv2


# take image input and assign variable to it
img = "4.jpeg"


# function to convert image into sketch
def rgb2gray(rgb):
    # 2 dimensional array to convert image to sketch
    return np.dot(rgb[..., :3], [0.2989, 0.5870, .1140])


def dodge(front, back):

    # if image is greater than 255 (which is not possible) it will convert it to 255
    final_sketch = front*255/(255-back)
    final_sketch[final_sketch > 255] = 255
    final_sketch[back == 255] = 255

    # to convert any suitable existing column to categorical type we will use aspect function
    # and uint8 is for 8-bit signed integer
    return final_sketch.astype('uint8')


ss = imageio.imread(img)
gray = rgb2gray(ss)

i = 255-gray


# to convert into a blur image
blur = scipy.ndimage.filters.gaussian_filter(i, sigma=13)


# calling the function
r = dodge(blur, gray)


cv2.imwrite('4.png', r)

Approach 2:

Import cv2:

--> pip install cv2

Then we will import cv2 inside our code, after that, we will use some of the following functions: 

1. imread()- This function will load the image i.e in the specified folder. 

2. cvtColor()- This function takes color as an argument and then changes the source image color into that color.

3. bitwise_not()- This function will help the image to keep the properties as same by providing the masking to it.

4. GaussianBlur()- This function is used to modify the image by sharpening the edges of the image, smoothen the image, and will minimize the

blurring property.

5. divide()- This function is used for the normalization of the image as it doesn’t lose its previous properties.

Finally will save the image using imwrite() function.

Python3

import cv2
image = cv2.imread('Image.jpg')  # loads an image from the specified file
# convert an image from one color space to another
grey_img = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
invert = cv2.bitwise_not(grey_img)  # helps in masking of the image
# sharp edges in images are smoothed while minimizing too much blurring
blur = cv2.GaussianBlur(invert, (21, 21), 0)
invertedblur = cv2.bitwise_not(blur)
sketch = cv2.divide(grey_img, invertedblur, scale=256.0)
cv2.imwrite("sketch.png", sketch)  # converted image is saved as mentioned name

Example 1:

Input image:

Output:

Example 2:

Input image:

Output: