Contrast Image
Introduction
Contrast image is the product of expanding the color distribution in the image.
and by doing so with the equalization on image's CDF, we will get more brighter / darker pixels
which will give us the contrast in the image.
lets explore that with Python & Open CV.
import cv2
import matplotlib.pyplot as plt
path_gorilla = "path/gorilla.jpg" # your image path
gorilla = cv2.imread(path_gorilla) # reading image
gorilla_grey = cv2.cvtColor(gorilla,cv2.COLOR_BGR2GRAY) # convert to Gray Scale
plt.figure(num=0,figsize=(12,6))
plt.imshow(gorilla_grey,cmap='gray')
plt.title('Gorilla Gray-Scaled Image')
histr = cv2.calcHist([gorilla_grey],[0],None,[256],[0,256])
plt.figure(num=1,figsize=(12,6))
plt.bar(list(range(0,256)),histr[:,0],align='center',color='b')
plt.xlim([0,256])
plt.xlabel('Pixels Value')
plt.ylabel('Frequency')
plt.title('Gray Scale Colors Distribution')
1. We can see that most pixels in the image are darker around 40.
2. We can see the high slops in ranges 0 ~ 40 and 170~200.
Now our gold is to reduce the slops and expand the distro for receive more contrast.
eq_gorilla = cv2.equalizeHist(gorilla_grey) # Equalization
histr_eq = cv2.calcHist([eq_gorilla],[0],None,[256],[0,256]) # new distro
plt.figure(num=3,figsize=(12,6))
plt.bar(list(range(0,256)),histr_eq[:,0],align='center',color='b')
plt.plot(histr,color='r')
plt.legend(['Before Equalization','After Equalization'])
plt.xlim([0,256])
plt.xlabel('Pixel Value')
plt.ylabel('Frequency')
plt.title('Gray Scale Image Distribution')
plt.figure(num=4,figsize=(12,6))
plt.imshow(eq_gorilla,cmap='gray')
1. We can see the distribution expand, the pick around 170 shift right around 230
which give more bright pixels in the new image.
2. We can see the slopes are more moderate all over the image.
The Result:
