NATIONAL INSTITUTE OF TECHNOLOGY, ROURKELA – 769008

END-SEMESTER EXAMINATION, 2018

SESSION: 2018-2019 (Autumn)
M. Tech. 1st Semester (ESC) + Dual Degree 9th Semester
Subjec Code: EE6131 & EE653 Subject Name: Digital Image Processing Dept. Code: EE
No. of pages : 2 Full Marks : 50 Duration: 3 hrs
Qs. Answer all questions. marks
Illustrate your answers with neat sketches as and when necessary.
1. a. For a 2-band sub-band coding applied to multi-resolution, it is found that the synthesis filters are 4
modulated versions of the analysis filters. Derive the same having the condition that defines
orthonormality for perfect reconstruction filter banks. Use Z-domain for the derivation.

b. Using the above in (a), compute and draw the coefficients of the synthesis filter g1(n), and 3
analysis filters h0(n), h1(n), given g0(n) for 0 ≤ n ≤ 3 as
n g0(n)
0 0.24
1 0.71
2 -0.02
3 -0.18

2. a. In multi resolution analysis, how scaling function coefficients and wavelet function coefficients 5
are related? Considering the set of expansion functions composed of integer translation and
binary scaling spanned in a subspace V, define the corresponding wavelet function W. Elaborate
their functionality and properties.
b.
The fast wavelet transform (FWT) is a computationally efficient implementation of DWT. Write 5
the expressions for the same. Compute the 2-D FWT with respect to Haar wavelets of the 2X2
image
 2 1
A 
5 3 
Draw the required filter banks and label all inputs and outputs with the appropriate arrays.

c. Draw the wavelet ψ3,3 for the Haar wavelet function. Write an expression for the same in terms of 3
Haar scaling functions.

3. a. Given a patch of size n×n represented as a matrix and given the orthonormal matrix U of size 3
n×n, explain how you will compute the 2D-DCT coefficients of the patch.

b. Image Compression standard JPEG supports different modes. Name them. Draw the block 3
diagram for Hierarchical JPEG and explain its advantages.

c. Walsh Hadamard Transform (WHT) is a computationally simpler transformation that is useful in 4

transform coding. Write the expression for the forward and inverse kernels. Hence create a 4-
point 2-D Walsh-Hadamard basis image WH13.
4.a. 3

Fig.1

Fig. 1 shows an image with an incomplete edge that divides the image into two regions. Propose a
combination of morphological image processing based methods to connect the edge and fill the
regions with binary labels. Clearly mention your steps.
b. 3
Prove the duality of morphological operations i) dilation and erosion, ii) opening and closing
c. Let u = [u1, u2, 1] T and v = [v1, v2, 1] T denote the homogeneous coordinates of points in planes
4
P and Q respectively. Matching point pairs are related by a matrix H as vi = Hui, where i = 1, 2, . .
. , n. A geometric transform from P to Q in the form of a translation by (x0, x1), then a rotation by
θ, and then a scaling by (s0, s1) is described by the matrix

 0.83 0.46 6.8

H   0.54 0.78 2.39
 0 0 1

Determine the values of (x0, x1, s0, s1, θ).

5.a.
A certain inspection application gathers images of parts as they travel along a conveyor belt. It is
4
necessary to sort the parts into two categories: parts with holes and parts without holes. An
example of an image that might be taken by the inspection camera is shown in Fig.2. Propose a
method to identify and locate the objects of each category in the image so that they can be picked
up by a robotic system and placed in different bins. Assume that the imaging system knows
where each image pixel is located on the conveyor belt at every point in time. Provide an
annotated flow chart of the algorithm you propose.

Fig. 2

b. Describe hypothesis testing based region merging method of segmentation, a statistical approach
3
to segment the adjacent regions based on their gray labels. Whether this approach is preferable
than edge based method? Justify your answer.
c. Nonmaximal suppression followed by hysteresis thresholding is an important task in Canny’s 3
edge detection method. Elaborate these steps with their advantages.