Python grayscaling rgb image with percentage

Question

I can find many codes that help to convert RGB image to a grayscale image. But none of them shows me how to grayscaling with adjustable percentage like CSS supports. https://developer.mozilla.org/en-US/docs/Web/CSS/filter-function/grayscale

img {
  filter: grayscale(25%);
}

Could you please guide me to build a function in python to return a grayscale image in rgb mode:

rgb_image = rgb_to_gray(rgb_img, percent = 0.25)

Thank you in advance!

Answer 1

Here is another way to do that in Python/OpenCV. This example is 75% gray (and 25% original).

Input:

import cv2
import numpy as np

# read image
img = cv2.imread("firefox.jpg")

# convert to grayscale
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

# make gray into 3 channels
gray = cv2.cvtColor(gray, cv2.COLOR_GRAY2BGR)

# blend gray with original
result = cv2.addWeighted(gray, 0.75, img, 0.25, 0)

# save results
cv2.imwrite('firefox_75pct_gray.jpg',result)

# show results
cv2.imshow('result', result)
cv2.waitKey(0)
cv2.destroyAllWindows()

Result:

Answer 2

The idea provided by @fmw42 is the best way. However, you cannot use the grayscaled image as is in the function cv2.addWeighted(). Both the images used within the function must of the same dimensions. Since a grayscale converted image is of single dimension, it will throw an error Sizes of input arguments do not match.

So first lets create a possible grayscale image in three dimensions, it can be done as follows:

img = cv2.imread(image_path)

# split the image into three channels
b,g,r = cv2.split(img) 

# Formula for converting to grayscale single channel
#0.2989 * R + 0.5870 * G + 0.1140 * B 

r= 0.2989 * r 
g = 0.5870 * g 
b = 0.1140 * b 

res = b+g+r
# grayscale single channel image
res = res.astype(np.uint8)

# to obtain 3-channel grayscale image repeat res across 3 channels
gray = cv2.merge((res, res, res))

Now as mentioned we use cv2.addWeighted() function as follows:

# the percentage of gray ranges from 0 to 1.0
for i in np.arange(0.0, 1.1, 0.1):
    modified_img = cv2.addWeighted(gray, i, img, 1.0-i, 0)
    cv2.imshow('Toned image', modified_img) 

Results

Percentage: 0

Percentage: 30

Percentage: 70

Percentage :100