Lapis design effect with opencv and python

Question

I want to get the result of the image further to the right, so I can only get to the middle image using a Canny filter and a Gaussian filter. Does anyone know how I can get this result?

import cv2
img = cv2.imread("foto.jpg", 0 )
cany = cv2.Canny(img,100,200)
cany = 255 - cany
cany = cv2.GaussianBlur(cany, (5,5),0)
cv2.imshow('resultado', cany)
cv2.waitKey(0)
cv2.destroyAllWindows()

Answer 1

import cv2

# abrindo a imagem em escala de cinza
img_gray = cv2.imread('wonder-woman.png', cv2.IMREAD_GRAYSCALE)

# calculando o inverso, 255 é branco 0 é preto e aplicando o blur
img_gray_inv = 255 - img_gray
img_blur = cv2.GaussianBlur(img_gray_inv, (21,21), 0, 0)

# fazemos o blend com o cv2.divide
img_blend = cv2.divide(img_gray, 255 - img_blur, scale = 256)

#salvamos
cv2.imwrite('wonder-woman-v1.png',img_blend)

1. Convert the color image to grayscale.
2. Invert the image in grayscale to get a negative.
3. Apply a Guassian Blur to the negative of step 2.
4. Combine the grayscale image of step 1 with the blurred negative of step 3 using a Color Dodge.

Source for the resolution you find in this book here.