SRM Institute of Science and Technology
SET A
College of Engineering and Technology
School of Computing
SRM Nagar, Kattankulathur – 603203, Chengalpattu District, Tamil Nadu
Academic Year:2025-2026 (ODD)
Test: FT2 Date: 18.09.2025
Course Code & Title: 21CSE454T & COMPUTER VISION Duration: 2periods
Year & Sem: IV year/ VII Sem Max. Marks: 50
Course Articulation Matrix:
Course P PO PO PSO PSO
PO PO PO PO PSO 3
Outcome PO1 PO2 PO3 PO4 PO5 O 10 12 1 2
6 7 8 11
9
- - - - - - - - - - 2 - -
CO1 2 1
3 1 - - - - - - - - - 2 - -
CO2 1
- 2 - - - - - - - - - 2 - -
CO3 1
- - 2 - - - - - - - 2 - -
CO4 2 -
1 - 1 2 - - - - - - - 2 - -
CO5 2
Part – A (10 x 1 = 10 Marks)
Instructions: Answer all
Q. Questions Mark BL CO PO PI
No s Code
Mean filter is mainly used for: 1 L1 1 1 1.1.1
a) Smoothing and noise removal
1 b) Edge detection
c) Skeletonization
d) Boundary tracing
Gaussian filter differs from mean filter because it: 1 L1 1 1 1.1.1
a) Weighs nearer pixels more
2 b) Gives equal weight to all pixels
c) Is not smoothing
d) Is binary only
3 Global thresholding uses: 1 L1 2 1 1.2.1
a) A single threshold for entire image
b) Multiple thresholds
c) Adaptive windows
d) Fourier coefficients
Texture refers to: 1 L1 1 1 1.1.1
a) Spatial arrangement of pixel intensities
4 b) Shape outline only
c) Noise in image
d) Compression artifacts
A common transform-based texture method is: 1 L1 2 2 2.3.1
5 a) Wavelet transform
b) Chain code
c) Region growing
d) Labeling
The Euclidean distance between two pixels (x1, 1 L1 1 2 2.3.2
y1) and (x2, y2) is:
6
a) |x1−x2| + |y1−y2|
b) max(|x1−x2|, |y1−y2|)
c) √((x1−x2)² + (y1−y2)²)
d) None of these
7 Skeletonization is also called: 1 L1 2 2 2.3.1
a) Morphological closing
b) Medial axis transform
c) Histogram equalization
d) Fourier analysis
8 Which is a practical use of thinning? 1 L1 2 2 2.3.1
a) Optical Character Recognition (OCR)
b) Histogram matching
c) Image compression
d) Filtering
9 Region descriptors are important in: 1 L1 2 2 2.3.2
a) Object classification
b) Histogram equalization
c) Filtering
d) Binarization
10 Active contours are also called: 1 L1 2 2 2.3.1
a) Snakes
b) Skeletons
c) Chain codes
d) Fourier descriptors
SRM Institute of Science and Technology
SET A
College of Engineering and Technology
School of Computing
SRM Nagar, Kattankulathur – 603203, Chengalpattu District, Tamil Nadu
Academic Year:2025-2026 (ODD)
Test: FT2 Date: 18.09.2025
Course Code & Title: 21CSE454T & COMPUTER VISION Duration: 2 periods
Year & Sem: IV year/ VII Sem Max. Marks: 50
SET-A
PART B (4*5=20 marks)
Answer Any Four Question
Q. Questions Marks BL CO PO PI
No Code
11 What is grayscale image and binary image. How to 5 2 1 1 1.4.1
covert grayscale image into binary image. With
example.
12 Compare the different thresholding methods (Global, 5 2 1 1 1.4.2
Adaptive, and Otsu’s) used in image processing.
13 What are centroidal profiles in image processing? 5 2 2 2 2.2.2
Explain how they are used for object shape analysis
with a neat diagram.
14 A satellite captures a binary image of a farmland where 5 3 2 2 2.3.1
1 represents regions of crop growth and 0 represents
barren land. Using 4-connectivity, agricultural analysts
need to identify distinct crop clusters. For the image:
0 1 1 0 1 0 1
0 1 1 0 1 0 1
0 0 1 1 0 0 0
1 1 0 0 1 1 1
Identify and tell how many crop clusters are there?
15 Find boundary length for the following diagram. 5 3 2 2 2.2.1
PART C (2*10=20 marks)
Answer Any Two
Q. Questions Marks BL CO PO PI
No Code
16 What are operators present in Edge detection method. 10 2 1 1 1.2.3
a) Explain any one operator with proper example.
Or
16 Explain statistical texture analysis using Gray Level Co- 10 2 1 1 1.3.2
b) occurrence Matrix (GLCM). List important texture
features derived from it.
17 Apply the concepts of Skeleton and Thinning to explain 10 2 2 2c 2.1.1
a) how character shapes can be simplified for recognition
in an OCR (Optical Character Recognition) system.
Discuss the role of each method with the help of a neat
sketch.
Or
17 A factory uses an imaging system to inspect electronic 10 3 2 2 2.3.1
b) components. The captured 7×7 binary image (1 =
object pixel, 0 = background) is given below:
0110000
0110110
0000100
1100000
0100111
0000100
0 000100
i)Label all connected objects using 8-connectivity.
ii)Compute the size (number of pixels) of each
labeled object.
iii)Apply size filtering with a threshold: keep only
objects with ≥ 4 pixels.
iv)Draw the final binary image after size filtering.