Ton Duc Thang University

Faculty of Information Technology

4 Exercises
Exercise 1: Write a function to transform an input image based on the following transformation:
1. Rotation with the different degrees (45, -45, 120, -120, 180, 270)
2. Scaling images with scale = 2, 1/2
3. Translation image with tx = 50, ty = 100
For each case, the image result is saved with jpg format

Hint:
• cv2.warpAffine(img,M,(w, h))
• cv2.getRotationMatrix2D(center, angle, scale)

Exercise 2: Write a function to perform perspective transformation for images on Google Drive of lab2.
1. With book.png, P = {(159.0, 125), (263, 183), (160, 336), (38, 234)}
2. With paper1.png, P = {(33, 97), (292, 111), (436, 457), (162, 577)}
3. With ex2.png, P = {(36, 62), (196, 19), (238, 102), (59, 163)}

Hint:
• cv2.getPerspectiveTransform(point1, point2)
• cv2.warpPerspective(I, M, (width, height))

Exercise 3: Write functions to convert an image from RGB color-space to gray with the following cases:
1. The lightness method averages the most prominent and least prominent colors: (max(R, G, B) +
min(R, G, B))/2
2. The average method simply averages the values: (R + G + B)/3.
3. The luminosity method is a more sophisticated version of the average method:
(a) 0.21R + 0.72G + 0.07B
(b) 0.299R + 0.7587G + 0.114B
Exercise 4: Write functions to convert an image from RGB color-space to CMY color-space and reverse
with the following equation:      
C 255 R
M  = 255 − G
Y 255 B

Exercise 5: Write functions to convert an image from RGB color-space to YCbCr and reverse which based
on the equation below:
 
219(0.299R + 0.587G + 0.114B)
  16 +
Y  255 
 
Cb = 128 + 224(−0.169R − 0.331G + 0.5B) 
 255 
Cr  
224(0.5R − 0.419G − 0.081B)
128 +
255

Anh H. Vo - vohoanganh@tdtu.edu.vn 8
 Ton Duc Thang University
Faculty of Information Technology

 
256(Y − 0.00093Cb + 0.114Cr)
  −16 +
R  219 
 
G =  −128 + 256(Y − 0.3437Cb − 0.71417Cr) 
 224 
B  
256(Y − 1.77216Cb − 0.00099Cr)
−128 +
224
Exercise 6: Write function to conversion between RGB and YIQ with the following equations
    
Y 0.299 0.587 0.114 R
 I  = 0.596 −0.274 −0.322 G
Q 0.211 −0.523 0.311 B
    
R 1.0 0.956 0.621 Y
G = 1.0 −0.272 −0.649  I 
B 1.0 −1.106 1.703 Q
Exercise 7: Write function to conversion between RGB and YUV with the following equations
    
Y 0.299 0.587 0.114 R
U  = −0.147 −0.289 0.436  G ,
V 0.615 −0.515 −0.100 B
    
R 1.0 0.0 1.140 Y
G = 1.0 −0.395 −0.581 U 
B 1.0 2.032 0.00 V

Exercise 8: Write functions to compare two images according to the following:

• Mean Squared Error (MSE) :
m−1 n−1
1 ��
M SE = [f (x, y) − g(x, y)]2 , a value of 0 for MSE indicates perfect similarity
m.n i=0 j=0
• Peak Signal to Noise Ratio (PSNR):
M2
P SN R = 10log10
M SE
where M is the maximum of the intensity values in an image.
Exercise 9: Write a program to evaluate the results between the transform images and the original image
from the previous exercises.
1. Using MSE
2. Using PSNR
Exercise 10: Write a program to compute the negative image of each gray image which converted in
exercise 3.
Exercise 11: Write a simple digital image processing program which performs the previous tasks (Exercise
1 - 9). It notes that the arguments are typed on command line by using argparse library

Hint:
import argparse
parser = argparse.ArgumentParser()
parser.add argument(’square’, help=’display a square of a given number’, type=int)
args = parser.parse args()
print(args.square**2)

Anh H. Vo - vohoanganh@tdtu.edu.vn 9