1. What is the different distance measures used for image? Compute the De, D4 and D8 distances
between two pixels p and q.
0 1 1 1q
p
1 0 0 0
1 1 1 1 q
1p 1 1 0
2. Define: Adjacency, Connected set, Region and Boundary. Consider the image shown below,
Let V= {1, 2}, compute the lengths of the shortest 4-, 8- and m-path between p and q. If a
particular path does not exist between these two pixels, explain why.
PO-a, l
2 4 2 1 2q
2 3 0 2 2
3 0 4 1 0
1 4 2 2 3
3 0 1p 2 3
3. Consider two image subsets S1 and S2 as shown in the following figure. For V = {1}, determine
whether these two subsets are (i) 4-adjacent, (ii) 8-adjacent, (iii) m-adjacent. If yes, find path
length between points p and q:
S1 S2
0 0 0 0 0 0 1 1(q)
0 0 1(p) 0 0 1 0 0
0 0 1 0 1 1 0 0
0 1 1 1 0 0 0 0
0 1 1 1 0 0 1 1
Unit II
1. Filter the following image using a 3X3 neighborhood averaging by assuming
(a) zero padding
(b) pixel replication
1 2 3 2
4 2 5 1
1 2 6 3
2 4 6 7
2. Consider the point transformation which maps the input image f(m,n) into an output image
g(m,n) through the transformation T. The input and output images are defined as follows:
(i) Input image f(m,n)=r, where 0r1
(ii) Output image g(m,n)=s=T(r)) , where 0s1
(iii) The transformation is given by T(r)=ar+b
Determine the values of a and b in term of mr and r2 . The mean and variance of r. Assume
ms = 0.5 and s2 = 0.25 r2 .
Nagar Yuwak Shikshan Sansthas
4. If all the pixels in an image are shuffled, will there be any change in the histogram? Justify
your image.
5. The images shown are quite different, but their histograms are identical. Suppose that each
image is blurred with a 3x3 smoothing mask.
7 6 5 4 7 6 5 4
7 6 5 4 6 7 4 5
7 6 5 4 5 4 7 6
7 6 5 4 4 5 6 7
0.5
2 6 7
I/P Gray Level r
14. Consider an 8 bit image given below which is to be displayed on a 3 bit display device. What is
the transformation function required. Obtained the output image using that transformation.
5 150 85 105
221 175 60 120
85 236 23 45
250 76 160 198
1. An image has gray level range from [25, 125]. It is desired the stretch the gray levels such a
way that the output image will have the gray level range from [0,255]. Suggest the
transformation function and expressed output image in term of input image.
2. Obtain the processed image using Max filter. (Note: Replicating Rows & Column).
7 4 0 1
6 5 2 3
4 2 7 5
0 1 6 2
3. For 4X4, 4 bit image, obtained the image by considering higher two bits only.
11 4 10 8
12 9 2 5
4 2 7 5
0 1 6 2
Nagar Yuwak Shikshan Sansthas
Show that the corresponding filter in the spatial domain has the form.
h( x, y ) = A 2 2 ( x2 + y 2 )
2 2
(
H (u, v) = u M ) (
)
2 2
+ vN
2 2
9. The input image of size 4X4 is passed through a linear shift-invariant system. Determine the
output image. The image and filter is given below. Assume zero padding of the original image.
12 10 8 4
8 14 6 9
5 9 13 8
14 5 7 9
Nagar Yuwak Shikshan Sansthas
10. What are the basic steps for filtering in frequency domain?
Unit III
1. What do you mean by image segmentation? Explain technique used for image segmentation.
2. Segment the given image using Sobel and prewit edge detector for threshold value T= 50% of
maximum value of gray level of the resultant sobel gradient image
7 2 3 2
2 5 5 1
3 5 5 0
1 4 3 7
3. Explain spilt and Merge technique for image segmentation and segment the images shown
below using spilt and Merge. Similarity criterion is absolute difference between seed and pixel
should be less than 3. Also draw its quadtree representation.
1 1 1 6
2 2 5 6
1 5 6 6
7 6 7 6
1 1 1 2 2 1 1 1
1 2 1 6 6 7 1
2 2 7 6 1 2 2 2
1 6 6 6 7 1 2 1
1 1 2 3 7 6 6 1
1 1 7 6 6 7 7 1
1 6 6 7 7 2 2 1
2 2 2 2 2 1 2 1
4. Suppose that an image has the gray level probability density functions shown below. Here p1(z)
corresponds to objects and p2(z) corresponds to the background . Assume that P1=P2 and find
the optimal threshold between object and background pixel.
1 p2(z)
p1(z)
Nagar Yuwak Shikshan Sansthas
5. Give the equation for first order and second order derivative. Draw the gray level profile and
find gradient and laplacian of the image strip shown below
5 5 4 3 2 1 0 0 0 6 0 0 0 0 1 3 1 0 0 0 0 7 7 7 7
6. Give approximation used for first order and second order derivative.
7. Derive the expression for Laplacian of Gaussian filter 2G ( x, y ) and show that its average
( x2 + y 2 )
value is zero. where G ( x, y ) = e 2 2
1 2 5 5
2 3 5 14
2 4 15 13
5 12 13 12
12. Explain region growing technique for image segmentation and segment the image shown below
using region growing. Use following predicate
(i) Absolute gray-level difference between pixel and the seed has to be less than 50
(ii) The pixel has to be 8-connected to at least one pixel in that region.
Nagar Yuwak Shikshan Sansthas
1 2 5 5 4 3 1 0
2 3 5 14 4 5 245 2
2 210 230 240 15 5 234 215
5 220 255 215 244 232 255 235
6 214 234 240 2 230 212 245
237 234 11 3 1 12 243 4
4 8 9 8 15 13 5 5
1 6 10 9 15 5 2 2
0 3 6 14 5 2 1 255
0 1 255 2 3 2 2 1
13. Discuss the global processing via the Hough transform. Use Hough transform to find straight
line given data points at (0,1), (1,1), (2,2) and 3,3)
14. Explain clearly the basic adaptive thresholding.
15. Write a short note on optimum thresholding and give the equations for A, B and C
16. Write short notes on the Edge linking
17. Explain the technique for detecting the three types of gray-level discontinuities in a digital
image, viz : points, lines and edges
Unit IV
2. Show that the Fourier transform of of the 2-D sine function f ( x, y ) = A sin(u0 x + v0 y ) is the pair
A u v u0 v
of conjugate impulses F (u , v) = j u 0 ,v 0
u + , v + 0 .Hint: Use the
2 2 2
2 2
continuous version of the Fourier transform and express the sine in term of exponentials.
3. Explain the algorithm of adaptive local noise reduction filter. Apply this filter of size 3X3 to
the image shown below which has been corrupted by Gaussian Noise with variance n2 = 5 .
7 4 0 1
6 5 2 3
4 2 7 5
0 1 6 2
4. Explain the algorithm of adaptive mean filter. Apply this filter of size 3X3 to the image shown
below. Smax=3.
7 4 0 1
6 5 2 3
4 2 7 5
0 1 6 2
5. Consider the problem of image blurring caused by planner motion in the x-direction and y
at bt
direction with x0 (t ) = and y0 (t ) = for time T, find the blurring function H (u , v ) . You
T T
may assume that shutter opening and closing times are negligible.
6. Obtain the equations for the bandpass filters corresponding to the band reject filters.
1 1 1 1
1 2 2 1
h(m, n) =
1 2 2 1
1 1 1 1
9. Give the expressions for Notch filter which rejects the frequencies in some neighborhood.
10. Problem No. 5.16 from Digital Image Processing 2nd and 3rd edition by Gonzalez and Woods.
Unit V
Prove that if an image f ( x, y ), 0 x M 1, 0 y N 1 is multiplied by checkerboard
1. pattern
M N
(1) x+ y ,
then its DFT is centered at 2 2
Generate NXN kernel matrix for following transformation : Discrete Fourier transform,
Discrete Cosine Transform, Discrete Sine Transform, Hadamard Transform, Haar Transform
for N=4 and N=8; Discrete Slant Transform for N=4. Verify their properties (Real,
2. Orthogonal and Symmetric). Also apply this 4X4 transformation on the image shown below
4 5 6 5
1 7 1 2
X =
4 2 0 2
6 4 2 1
4. Prove that cosine transform (DCT) is not the real part of Discrete Fourier Transform
5. Prove that real and imaginary parts of DFT unitary matrix are not orthogonal.
For the 2x2 transform A and the image U calculates the transformed image V and the basis
images.
7.
1 3 1 2 3
A= ; U =
2 1 3 1 2
4 2
11. X =
Find DFT ,DCT,, DST, Haar, Hadamard and slant transform of 1 3
1 3 1 1 1 1
12. A1 = ; A2 =
2 1 3 2 1 1
Check unitary and orthogonal property of
For the image shown in fig below find the value at F(0,0), where F(0,0) denotes the frequency
component at (0.0).
1 3 5 7
0 2 4 6
6 7 7 6
4 6 7 6
Nagar Yuwak Shikshan Sansthas
Unit VI
1. What are different types of data redundancies present in images?
2. Explain coding redundancy, interpixel redundancy and psycho visual redundancy.
3. Explain the mean of quantifying the nature of loss. (Subjective and Objective fidelity criterion.)
4. Explain the functional block diagram of general image compression system.
5. Generate the code to transmit the message a1a2 a3 a3 a4 using arithmetic technique. The symbols are
having probability as shown below
SoureSymbol Probabilty
a1 0.2
a2 0.2
a3 0.4
a4 0.2
6. Explain zonal coding and threshold coding implementation.
7. Explain lossless predictive coding system.
8. Explain lossy predictive coding system.
9. Decode the message 0.23355 using arithmetic coding.
Symbol Probablity
a 0.2
e 0.3
i 0.1
o 0.2
u 0.1
! 0.1
10. For the following symbol with probabilities
Symbol Probablity
a1 0.1
a2 0.4
a3 0.06
a4 0.1
a5 0.04
a6 0.3
Nagar Yuwak Shikshan Sansthas
quantized with 4-bit accuracy, compute the rms error and rms signal-to-noise ratios for the
quantized data.
12. Consider the simple 4 8 , 8-bit image
21 21 21 95 169 243 243 243
21 21 21 95 169 243 243 243
21 21 21 95 169 243 243 243
21 21 21 95 169 243 243 243
(a) Compute the entropy of the image
(b) Compress the image using Huffman coding
(c) Compute the compression achieved.
13. Given a four symbol source {a, b, c, d } with source probability {0.1, 0.4, 0.3, 0.2} , arithmetically