Question

We’re sure you've seen old black-and-white (actually, grayscale) photos that, over time, have gained a
yellowish tint. We can mimic this effect by creating sepia-tinted images.
There are several different ways to sepia-tint an image. In one of the simplest techniques, the image is
first converted to grayscale, and then each pixel's red and blue components are adjusted, leaving the
green component unchanged. Here's the algorithm:
· First, we convert the image to grayscale, because old photographic prints were grayscale. After this
conversion, each pixel is a shade of gray, so its red, green and blue components are equal.
· Next, we tint each pixel so that it's a bit yellow. In the RGB system, yellow is a mixture of red and
green. To make a pixel appear slightly more yellow, we can simply decrease its blue component by a
small percentage. If we also increase the red component by the same percentage, the brightness of
the tinted pixel is essentially unchanged. The amount by which we change a pixel's red and blue
components depends on whether the pixel is a dark gray, a medium gray, or a light gray.
o If the pixel's RGB values are less than 63, the pixel is in a shadowed area (it's a dark gray), so
the blue component is decreased by multiplying it by 0.9 and the red component is increased
by multiplying it by 1.1. (Note that we only have to compare one of the pixel’s components
to 63 – not all three – because all three components in a shade of gray have the same value.)
o If the pixel's RGB components are between 63 and 191 inclusive, the pixel is a medium gray.
The blue component is decreased by multiplying it by 0.85 and the red component is
increased by multiplying it by 1.15.
o If pixel's RGB components are greater than 191, the pixel is in a highlighted area (it's a light
gray), so the blue component is decreased by multiplying it by 0.93 and the red component is
increased by multiplying it by 1.08.
Develop a filter named sepia that returns a copy of the input image in which all the pixels have been
sepia-tinted. Don’t implement the grayscale algorithm in this function. Instead, import the grayscale
function from simple_Cimpl_filters.py,

Answer 1

please refer to below attached photos for clear understanding and implementation of the problem :

Note : please replace the load_gray function with the grayscale fucntion that u currently have.

Activities Atom - Thu 11:05 PM tint_image.py - n/yogeesh/random_stuff – Atom File Edit View Selection Find Packages Help tint_image.py х dog.jpg 1 2 import numpy as np import matplotlib.pyplot as plt import matplotlib.image as img import cv2 الا 4 Louille 5 6 7 # create a numpy array of same shape as image filled with 0 # get the pixel for each height and width. # get the increase and decrease factor for red and blue channels. # as cv2 format is BGR , store the pixel * decrease_factor in first channels # store same pixel in second as green is unchanged. # store the pixel * increase_factor in last channel for red. # return the new image. def tint_image(image): print(image. shape) tinted_image = np.zeros_like(image) for h in range(image. shape[0]): for w in range(image.shape[1]): pixel = image[h,w, 0] if pixel < 63: decrease factor = 0.9 increase_factor = 1.1 EN a 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 elif pixel >= 63 and pixel < 191: decrease factor = 0.85 increase_factor = 1.15 else: decrease factor = 0.93 increase_factor = 1.08 tinted_image[h,w, 0] = int(pixel * decrease_factor) tinted_image[h,w, 1] = pixel tinted_image[h,w, 2] = int(pixel * increase_factor) return tinted_image tint_image.py 43:1 UTF-8 Python GitHub Git (0)

Activities Atom - Thu 11:06 PM tint_image.py - -/yogeesh/random_stuff – Atom File Edit View Selection Find Packages Help tint_image.py X dog.jpg 20 21 22 decrease factor = 0.9 increase factor = 1.1 E elif pixel >= 63 and pixel < 191: decrease factor = 0.85 increase factor = 1.15 23 24 25 26 27 28 29 else: decrease factor = 0.93 increase_factor = 1.08 30 31 32 33 34 35 tinted_image[h,w, 0] int(pixel * decrease factor) tinted_image[h,w,1] = pixel tinted_image[h,w, 2] = int(pixel * increase_factor) return tinted_image 36 a # this fucntion loads a grayscale image using cv2 and convert grayscale to three channel grayscale. # please replace this function with the one you have for grayscale fucntion def load_gray(filename) : gray = CV2.imread(filename, 0) three_channel_gray = cv2.cvtColor(gray, CV2.COLOR_GRAY2BGR) return three_channel_gray 3 37 38 39 40 41 42 43 44 45 46 47 48 def main(): three_channel_gray = load_gray("dog.jpg") tinted_image = tint_image(three_channel_gray) tinted_image) 49 CV2.imshow("tinted_image" CV2.waitkey (0) CV2.destroyAllWindows (). 50 51 52 53 54 if main ": name main(). tint_image.py 43:1 UTF-8 Python GitHub Git (0)

Activities Thu 11:06 PM OX dog.jpg - -/yogeesh/random_stuff - Atom File Edit View Selection Find Packages Help yogeesh@yogeesh-Latitude-3400: ~/yogeesh/random_stuff File Edit View Search Terminal Help (yenv) yogeesh@yogeesh-Latitude - 3400:-/yogeesh/random_stuff$ python tint_image.py fyenu) tint_image.py X Ridog.jpg X Auto tinted_image EN a (x=197.v=0)~R:64 G:59 B:53 dog.jpg 453x340 31.32KB GitHub Git (0)

code :

import numpy as np
import matplotlib.pyplot as plt
import matplotlib.image as img
import cv2

# create a numpy array of same shape as image filled with 0
# get the pixel for each height and width.
# get the increase and decrease factor for red and blue channels.
# as cv2 format is BGR , store the pixel * decrease_factor in first channels
# store same pixel in second as green is unchanged.
# store the pixel * increase_factor in last channel for red.
# return the new image.
def tint_image(image):
tinted_image = np.zeros_like(image)
for h in range(image.shape[0]):
for w in range(image.shape[1]):
pixel = image[h,w,0]
if pixel < 63:
decrease_factor = 0.9
increase_factor = 1.1

elif pixel >= 63 and pixel < 191:
decrease_factor = 0.85
increase_factor = 1.15

else:
decrease_factor = 0.93
increase_factor = 1.08

tinted_image[h,w,0] = int(pixel * decrease_factor)
tinted_image[h,w,1] = pixel
tinted_image[h,w,2] = int(pixel * increase_factor)

return tinted_image

# this fucntion loads a grayscale image using cv2 and convert grayscale to three channel grayscale.
# please replace this function with the one you have for grayscale fucntion .
def load_gray(filename):
gray = cv2.imread(filename , 0)
three_channel_gray = cv2.cvtColor(gray , cv2.COLOR_GRAY2BGR)
return three_channel_gray

def main():
three_channel_gray = load_gray("dog.jpg")
tinted_image = tint_image(three_channel_gray)

cv2.imshow("tinted_image" , tinted_image)
cv2.waitKey(0)
cv2.destroyAllWindows()

if __name__ == "__main__":
main()