ELECTRONICS & COMMUNICATION ENGINEERING

TOPICS IN COMPUTER CATEGORY L T P CREDIT

ECT386
VISION VAC 3 1 0 4

Preamble: This course aims to develop the knowledge of various methods, algorithms and
applications of computer vision

Prerequisite: Nil

Course Outcomes: After the completion of the course the student will be able to

Apply basic point operators and 2D transforms for digital filtering

CO1 operations
Apply various algorithms for morphological operations and binary shape
CO 2 analysis.
Understand the theoretical aspects of image formation and 3D camera
CO3 models and vision system.

Apply edge, corner detection methods and optical flow algorithms to locate
CO 4 objects in an image/video.

Analyse 3D images and motion of objects in a given scene using appropriate

CO5 algorithms computer vision algorithms for real time practical applications.

Mapping of course outcomes with program outcomes

PO PO PO PO PO PO PO PO PO PO 10 PO PO
1 2 3 4 5 6 7 8 9 11 12
CO 1 3 3 2 2 2 2 3
CO 2 3 3 2 2 2 2 3
CO 3 3 3 3 2 2 2 3
CO 4 3 3 3 2 2 2 3
CO 5 3 3 3 2 2 2 3

Assessment Pattern

Bloom’s Category Continuous Assessment End Semester

Tests Examination
1 2
Remember K1 10 10 10
Understand K2 10 10 20
Apply K3 20 20 70
Analyse K4 10 10
Evaluate
Create

Mark distribution

Total Marks CIE ESE ESE Duration

150 50 100 3 hours

 ELECTRONICS & COMMUNICATION ENGINEERING

Continuous Internal Evaluation Pattern:

Attendance : 10 marks
Continuous Assessment Test (2 numbers) : 25 marks
Assignment/Quiz/Course project : 15 marks

End Semester Examination Pattern: There will be two parts; Part A and Part B. Part A
contain 10 questions with 2 questions from each module, having 3 marks for each question.
Students should answer all questions. Part B contains 2 questions from each module of which
student should answer any one. Each question can have maximum 2 sub-divisions and carry
14 marks.

Course Level Assessment Questions

Course Outcome 1 (CO1): Apply basic point operators and 2D transforms for digital
filtering operations
1. Why histogram transformations are applied in a grey scale image and what output is
observed in that case.
2. Find filtered image using LP/HP/Smoothing/Median filter.
3. Describe the working principle of Homomorphic filter.

Course Outcome 2 (CO2): Apply various algorithms for morphological operations and
binary shape analysis
1. List different morphological operators and describe about each one in detail.
2. To describe connected component labelling and to apply it in a given image pixel set.
3. Find 8-point connectivity and Chain code of a given image pixel diagram.

Course Outcome 3 (CO3): Understand the theoretical aspects of image formation

and 3D camera models and vision system.

1. Differentiate between Monocular and binocular imaging system.

2. Compare and Contrast Orthographic and Perspective Projection methods.
3. Describe image formation and geometric transformations in 3D Camera Models.
4. Write short notes on 3D-Imaging system.
5. Briefly describe a stereo vision system.

Course Outcome 4 (CO4): Apply edge, corner detection methods and optical flow
algorithms to locate objects in an image/video.

1. What is the role of edge detection and corner detection in Computer Vision
applications?
2. Describe Canny’s edge detection algorithm.
3. Mention the steps in Harris corner detection algorithm and explain how it is employed
to detect corners in an image.
4. State with necessary mathematical steps, how Hough transform is employed for
detecting lines and curves in detecting an image.
 ELECTRONICS & COMMUNICATION ENGINEERING
Course Outcome 5 (CO5): Analyse 3D images and motion of objects in a given scene using
appropriate computer vision algorithms for real time practical applications.
1. Find Eigen values and Eigen Vectors of the following matrix

[ ]
8 −8 − 2
A= 4 −3 − 2
3 −4 1

2. Analyse a given video to track a moving object in it.

3. To detect a particular object from the background.
4. To detect signboards/ pedestrian crossings/pedestrians from a moving vehicle.
5. To classify/segment a particular set of image using CV algorithms.
6. Analyse a given image/video using Machine learning/Deep learning algorithms.
7. Use trained networks to analyse a video using ML algorithms.
8. To use Deep neural networks/CNN/YOLOvx, to analyse images/videos

SYLLABUS

Module 1
Review of image processing techniques: Filtering, Point operators-Histogram Based
operators, neighbourhood operators, Thresholding - linear filtering – development of filtering
masks - 2D Fourier transforms – filtering in frequency domain, Homomorphic filtering

Module 2
Mathematical Operators: Binary shape analysis: Basics of Morphological operations,
structuring element, Erosion, Dilation, Opening and Closing, Hit-or-Miss Transform,
Connectedness, object labelling and counting , Boundary descriptors – Chain codes.

Module 3
Camera models: Monocular and binocular imaging system, Orthographic and Perspective
Projection, Image formation, geometric transformations, Camera Models (Basic idea only),
3D-Imaging system-Stereo Vision.
Module 4
Feature Detection: Edge detection – edges, lines, active contours, Split and merge, Mean
shift and mode finding, Normalized cuts, Graph cuts, energy-based and Canny’s methods.
Corner detection, Harris corner detection algorithm, Line and curve detection, Hough
transform
SIFT operators, Shape from X, Shape Matching, Structure from motion.

Module 5
Motion Analysis- Regularization theory, Optical Flow: brightness constancy equation,
aperture problem, Horn-Shunck method, Lucas-Kanade method. (Analysis not required)
Object Detection and Object classification: SVM, Linear discriminant analysis, Bayes rule,
 ELECTRONICS & COMMUNICATION ENGINEERING
ML.
Face detection, Face Recognition, Eigen faces, 3D face models
Applications of Computer Vision: Context and scene understanding, Real Time applications:
Locating road way and road marking, locating road signs and pedestrians.

Text Books
1. E. R .Davies, Computer and Machine Vision -Theory Algorithm and Practicalities,
Academic Press, 2012
2. Richard Szeliski, Computer Vision: Algorithms and Applications, ISBN 978-1-
84882-935-0, Springer 2011.
3. David Forsyth and Jean Ponce, Computer Vision: A Modern Approach, Pearson India,
2002.

Reference Books
1. Goodfellow, Bengio, and Courville, Deep Learning, MIT Press, 2006.
2. Daniel Lelis Baggio, Khvedchenia Ievgen, Shervin Emam, David Millan Escriva,
Naureen Mahmoo, Jason Saragi, Roy Shilkrot, Mastering Open CV with Practical
Computer Vision Projects, Packt Publishing Limited, 2012
3. Simon J D Prince, Computer Vision: Models, Learning, and Inference, Cambridge
University Press, 2012.
4. R. J. Schalkoff, Digital Image Processing and Computer Vision, John Wiley, 2004.

Course Contents and Lecture Schedule

No. Topic No. of

Lectures
1 Introduction
1.1 Review of image processing techniques: filtering. 2
1.2 Point operators- Histogram, neighbourhood operators, thresholding– 3
development of filtering masks
1.3 2D Fourier transforms – filtering in frequency domain, 3
homomorphic filtering
2 Mathematical Operators:
2.1 Basics of Morphological operations , structuring element 2
2.2 Binary shape analysis : Erosion, Dilation, Opening and 2
Closing, Hit-or-Miss Transform
2.3 Connectedness, object labelling and counting 2
2.4 Boundary descriptors –Chain Codes. 2
3 Camera models
3.1 Monocular and binocular imaging system 2
3.2 Orthographic & Perspective Projection 2
3.3 Image formation, geometric transformations, camera Models(Basic 3
idea only)
3.4 3D-Imaging system-Stereo Vision 2
4 Feature Detection:
4.1 Edge detection – edges, lines, active contours, Split and merge, Mean 4
shift and mode finding, Normalized cuts, Graph cuts, energy-based
 ELECTRONICS & COMMUNICATION ENGINEERING
and Canny’s methods.
4.2 Corner detection, Harris corner detection algorithm, Line and curve 3
detection, Hough transform
4.3 SIFT operators, Shape from X, Shape Matching 3
5 Motion Analysis
5.1 Motion Analysis- Regularization theory, Optical Flow: brightness 3
constancy equation, aperture problem,
Horn-Shunck method, Lucas-Kanade method (Analysis not required)
5.2 Object Detection and Object classification: SVM, Linear discriminant 4
analysis, Bayes rule, maximum likelihood, Face detection, Face
Recognition, Eigen
faces, 3D face models
5.3 Applications of Computer Vision: Context and scene understanding, 3
Real Time applications:
Locating road way and road marking, locating road
signs and pedestrians
 Simulation Assignments
ELECTRONICS & COMMUNICATION ENGINEERING
The following simulations can be done in OpenCV/SciLab/ MatLab

1. Design and implementation of basic digital filters.

2. Apply thresholding operations in a digital image.
3. Apply point operators in an image – averaging/smoothing, 2D - masks(3 types),

4. Apply morphological operations in a selected image like fingerprint/ archaeological

scripts.
5. Implement filters in 2D-frequency domain using Gaussian/Homomorphic filters in a
particular satellite image or forensic image.
6. Write algorithms for connected component labelling in a given image pixel set.
7. Detect a coin/ball against the background using background subtraction and with
appropriate edge detection algorithms.
8. Locate corners of a particular image like boxes/ building/TV screen etc
9. Write a program to implement brightness constancy equation.
10. Analyse the optical flow of a given video using Horn-Schunk method or/and Lucas-
Kannade method/s.
11. Use PCA for dimensionality reduction in detecting faces using Eigen values.
12. Implement SVM/LDA for a practical application.
13. Create an attendance system by implementing face recognition method, among a set
of students.
14. With OpenCV library, implement real time scene analysis for traffic regulation.
(cases such as detecting road signs/ pedestrians/track a particular vehicle/ detect
traffic lights/detect number plate of a vehicle/ detect accidents/ accident scene
analysis etc., etc.).
15. Use ML/DL algorithms to implement object detection/identification/classification,
with trained neural networks for applications in medical/agricultural/sports fields.
Model Question paper
ELECTRONICS & COMMUNICATION ENGINEERING
APJ ABDUL KALAM TECHNOLOGICAL UNIVERSITY
SIXTH SEMESTER B.TECH DEGREE EXAMINATION, (Model Question Paper)
Course Code: EC386

Course Name: TOPICS IN COMPUTER VISION

(Minor)
Max. Marks: 100 Duration: 3 Hours
PART A

Answer ALL Questions. Each carries 3 marks.

1 Compare and Contrast DoG and LoG. K1

2 An image is convolved with the matrix given below. Express analytically the K2
relation between original and modified image.

3 List any three computer vision applications of object labeling and counting. K2

4 Describe steps in identifying connected components in 4-connectivity and 8- K2

connectivity cases.

5 Describe the working principle of pin- hole camera K2

6 Compare and contrast perspective and orthographic projection. K2

7 Mention the concept of identifying structure from motion. K2

8 Illustrate how graph cut method is employed for edge detection K2

9 Define eigen values and eigen vectors. K2

10 Differentiate between SVM and LDA. K2

PART – B

Answer one question from each module; each question carries 14 marks.

Module - I

11 a What is thresholding? Briefly describe different methods of thresholding. (6)

CO1

K1
11 b. Apply 2D-DFT on the given image pixel values (8)
ELECTRONICS & COMMUNICATION ENGINEERING
CO1
K3

OR

12 a (8)
Describe how histogram transformations are applied on a grey scale image and CO1
explain what output is observed in each case. K3

12 b. Describe the working principle of Homomorphic filter, with different (6)

mathematical steps involved. CO1
K2

Module - II

13a An image A, and a structuring element B are given. Apply B on A, to find (8)
resultant images after the dilation and erosion process. Origin is given as ‘O’; and
note that it is not part of the structuring element.
CO2
K3

1 1 1

1 1
A=
1 1 1

1 1

1 O 1
B =

13 b. Using Hoshen–Kopelman algorithm, assign connected component labelling for the (6)
given image in the diagram below.

CO2
K3
 ELECTRONICS
OR & COMMUNICATION ENGINEERING

14a Apply Hit-or-Miss transform on the given binary image to detect right angle (8)
convex corners from left-top to right-bottom.

CO2
K3

14b. (6)

x x x CO2
x x K3
x x

x x

x x

For the given image grid, find out the 8-chain Freeman Code. Also write down the
chain number of the code.

15 Describe image formation and geometric transformations in 3D Camera Models. (14)

CO3

K2

OR

16a Briefly describe a stereo vision system. (7)

CO3
K2

16 b Elucidate on 3D-Imaging Camera system. (7)

CO3
K2
 Module - IV
ELECTRONICS & COMMUNICATION ENGINEERING
17a With the help of a general algorithm, narrate how Hough Transform can be (6)
employed for interest point detection. CO4
K3

17 b Unveil different steps involved in Harris corner detection algorithm and explain (8)
how it is employed to detect corners in an image
CO4
OR
K3

18 a. Apply Sobel and Prewitt masks on the given image grids and compute Gx and Gy (6)
gradients of the images. CO4
K3

18 b. Referring to Canny’s algorithm, describe how it is employed to detect edges of an (8)

image.
CO4
K3

Module - V

19 a Derive brightness constancy equation. (6)

CO5
K3

19 b Describe with algorithmic steps, the Horn-Shunk method used for the estimation (8)
of optical flow. CO5
K3

20 a Illustrate the concept of dimensionality reduction using Principal Component (7)

Analysis (PCA) with all mathematical steps involved. CO5

K3

20b. With the help of a flow chart, illustrate an in-vehicle vision system, for locating (7)
roadways and pedestrians, as a practical application of computer vision. CO5

K3