Question

What to do

This assignment is going to be pretty simple - we're just going compute frequency counts of certain colors in certain ranges (Histogram of the image).

Because there is no natural way of segregating RGB color ranges, we usually do this color by color. So you'll be computing frequency counts for individual color components individually.

The expectation is that you will write a function as described by a python function signature below (while you are free to do whatever you want, I think doing future parts of the assignment will be easier if the code is structured this way) -

Think about what happens when one is comparing images of different sizes -

• Lets say A is an 4 by 4 image where half of the colors are gray (rgb(128, 128, 128)), and the other half is white (rgb(255, 255, 255))
• When you compute frequencies for Red component, you'll find that 128 occurs 8 times and 255 occurs 8 times also.
• Lets say B is an 8 by 8 image where half of the colors are gray (rgb(128, 128, 128)), and the other half is white (rgb(255, 255, 255))
• When you compute frequencies for Red component, you'll find that 128 occurs 32 times and 255 occurs 32 times also.
• This would mean that images that are basically the same in terms of colors (both half gray and white), end up with different histograms.

To make sure similar images have similar histograms regardless of image size, it is good to normalise frequencies by dividing each frequency count by the total number of pixel in the image.

• After normalization, frequency for red 128 in A would be 8 / (4 * 4) = 0.5
• After normalization, frequency for red 128 in B would be 32 / (8 * 8) = 0.5
• You will be given an image as a text file in the format we've been using so far - assignment_files.zip

Answer 1

# PLEASE LIKE THE SOLUTION
# FEEL FREE TO DISCUSS IN COMMENT SECTION

# Python Program

# Assure that matplot and cv2 are installed in your ide used for execution of this program
# if you not require plot of histogram you may remove optional code mention below and remove import matplotlib.pyplot as plt
import matplotlib.pyplot as plt
import cv2

# this function calulates frequency and draw histogram and print in terminal
def draw_histogram(image):
   # initilaise the key value pair
   freqForRed = {}
   freqForGreen = {}
   freqForBlue = {}
  
   # get height of image
height = image.shape[0]
   width = image.shape[1]

   # initilaising with 0 for all 0 to 255 pixel frequency
   for i in range(256):
       freqForRed.update({i : 0})
       freqForGreen.update({i : 0})
       freqForBlue.update({i : 0})

   # for each pixel get bgr value and update frequency
  
   for i in range(0,height):
       for j in range(0,width):
           color = image[i][j]
           blueValue = int(color[0])
           greenValue = int(color[1])
           redValue = int(color[2])
          
           # update red frequency for this pixel
        value = freqForBlue[blueValue]
           freqForBlue.update({freqForBlue[blueValue] : value + 1})

           value = freqForGreen[greenValue]
           freqForGreen.update({freqForGreen[greenValue] : value + 1})

           value = freqForRed[redValue]
           freqForRed.update({freqForRed[redValue] : value + 1})

   redList = []
   blueList = []
   greenList = []

   # now normalising by dividing by its size
   size = float((height * width))
   for i in range(256):
       value = freqForRed[i];
       redList.append((value/size))
      
       value = freqForGreen[i];
       blueList.append((value/size))

       value = freqForBlue[i]
       greenList.append((value/size))

   #print key and normalised frequency
   for i in range(256):
       print("Normalized Frequency Red"+str(i)+" is "+str(redList[i])+" Green"+str(i)+" is "+str(greenList[i])+" Blue"+str(i)+" is "+str(blueList[i]))

   # OPTIONAL CODE FOR PLOTING
   # PLOT RED
x = list(freqForRed.keys())
   plt.plot(x, redList)
   plt.gcf().canvas.set_window_title("RED")
   plt.show()

# PLOT GREEN
   plt.plot(x, greenList)
   plt.gcf().canvas.set_window_title("GREEN")
   plt.show()

   # PLOT BLUE
   plt.plot(x, blueList)
   plt.gcf().canvas.set_window_title("BLUE")
   plt.show()

img = cv2.imread('./image.jpg')
draw_histogram(img)

# SAMPLE OUTPUT

# HISTOGRAM