Computer Vision

Introduction

Dr. Surya Prakash

[email protected]

CS419/619: Computer Vision

 Outline
 Course Introduction

 Tentative Topics to be Covered

 Text Books/References

 Evaluation Policy

2 CS419/619: Computer Vision

 Introduction
 Computer Vision = Computer + Vision

– Vision – state of being able to see

– Computer Vision
• enables computers to see (capture images, videos, etc.)
and analyze the world

3 CS419/619: Computer Vision

 Introduction
 Human Vision
– Through our vision, we see the world (through eyes),
analyze it (through our brain), and take necessary actions
based on derived information
– How does human visualize/understand the real world

 Computer Vision
– Understanding the real world with the help of computer
 See the real world (capture images, videos, etc)
 derive meaningful information from the images, videos and, other
visual inputs
 and take actions based on the derived information

4 CS419/619: Computer Vision

 Introduction
 Human Vision
–It can do amazing things like:
• Recognize people and objects
• Navigate through obstacles
• Understand mood in the scene
• Imagine stories

5 CS419/619: Computer Vision

 Introduction

 Computer Vision:
–Main task involves intelligent
interpretation of pictures or scenes
• Understanding the input image/ scene
• Extracting the information from the scene
• Description of the scene information

6 CS419/619: Computer Vision

 Components of a Computer Vision System

Camera

Lighting

Computer

Scene

Scene Interpretation

7 CS419/619: Computer Vision

 Components of a Computer Vision System

Computer Images/Scenes
Images/Scenes
Vision representation

Visualization

8 CS419/619: Computer Vision

 Human Vision Vs. Computer Vision

What a human sees

What a computer sees

9 CS419/619: Computer Vision

 Computer Vision – Selection of Techniques

 Techniques used to solve a particular problem:

– depends on the application domain
– nature of data being considered as input
Two different Problems/Applications:
 Face recognition
 Robot navigation Face
recognition

 Choice of techniques/features
Robot
– are problem specific navigation
– are chosen after some amount of trial & error
– experience helps in choosing techniques/features
Images: https://mobile.aivanet.com/2018/03/samsungs-galaxy-s10-rumored-to-feature-3d-facial-recognition-like-face-id-on-iphone-x/
https://boredomprojects.net/index.php/projects/robot-navigation-using-stereo-vision

10 CS419/619: Computer Vision

 Computer Vision–Selection of Techniques (Examples)
 Example 1:
– Face Recognition Vs. Robot Navigation
 Example 2:
– Analysis of remote sensed images Vs. Indoor
outdoor images
• Analysis of remote sensed images (satellite imagery)
may involve techniques of texture analysis
• However, these techniques may not be helpful in
analyzing images of indoor/outdoor scenes
 Example 3:
– Character Recognition Vs. Speech Recognition

11 CS419/619: Computer Vision

 Computer Vision Tasks
 Image representation
 Image description
 Feature extraction
 Image matching
 Stereo image processing
 Range data processing
 Image segmentation

12 CS419/619: Computer Vision

 Application Areas of Computer Vision

 Character Recognition
 Document Processing
 Biometric Applications
– verification of persons using biometric characteristics
– Biometric characteristics: face, iris & fingerprint
 Forensic Analysis
 Pose and Gesture Identification
 Automatic Inspection of Industrial Products
 Industrial Process Monitoring

13 CS419/619: Computer Vision

 Application Areas of Computer Vision…contd.

 Biomedical Engineering
– Diagnosis and Surgery

 Military surveillance and target identification

 Navigation and mobility

– for robots and unmanned vehicles - land, air and
underwater

 Remote sensing (using satellite imagery)

14 CS419/619: Computer Vision
 Application Areas of Computer Vision…contd.

 Geographic Information Systems (GIS)

 Safety and security (night vision)

 Traffic monitoring

 Sports (training and incident analysis)

 Entertainment and virtual reality

15 CS419/619: Computer Vision

 Computer Vision Vs. Image Processing
 In Computer Vision
– Input: Images, scenes
– Output:
• Intelligent interpretation of pictures or scenes
• Output may be represented in the form of scalar,
vector or matrix

16 CS419/619: Computer Vision

 Computer Vision Vs. Image Processing

 Image processing
– Input: Image (an array of pixels)
– Output: Image (another transformed array of
pixels )
– This transformation in some way represents an
improvement to the original array

Image Processing
Transformation

Input output
17 CS419/619: Computer Vision
 Computer Vision Vs. Image Processing
 IP deals with improving the pictorial
information
– improves the contrast of the image
– removes noise
– removes blurring caused by movement of
the camera during image acquisition
– it may correct geometrical distortions
caused by the lens

18 CS419/619: Computer Vision

 Computer Vision Vs. Image Processing
Image Processing

CV
IP

Original FACIAL
Image RETOUCHING

CV
IP

Images: https://hal.inria.fr/hal-02378939/document

19 CS419/619: Computer Vision

 Computer Vision Vs. Machine/Deep Learning
Gender Classification

Classifier

Image source: Chen, Cunjian, "Gender Classification from Facial Images" (2011). Graduate Theses, Dissertations, and Problem
Reports. 3449. https://researchrepository.wvu.edu/etd/3449

20 CS419/619: Computer Vision

 Courses Related to Computer Vision
 It uses the concepts and techniques from
– Image Processing
– Artificial Intelligence
– Pattern Recognition
– Machine Learning
– Deep Learning
– Artificial Neural Networks
– Computer Graphics
– Optimization Techniques
– Fuzzy Theory
– VLSI and Computer Architecture
21 CS419/619: Computer Vision
 Topics Covered in this Course
 Image Processing
– Histogram Processing
– Fourier Transform
– Convolution and Correlation
– Image Enhancement
– Edge Detection
– Morphological Operations
 Hough Transform
 Statistical Techniques
– Least Square Fitting
– Principal Component Analysis

22 CS419/619: Computer Vision

 Topics Covered in this Course
 Feature detection and representation
– SIFT and SURF
 Segmentation
– Texture Analysis and Segmentation
 Optical Flow
 Object Segmentation
– Active Contour
 Object Recognition
 Pattern Classification

23 CS419/619: Computer Vision

 Pre-requisite
• Mathematics:
• Basic Linear Algebra [Matrix operations]
• Probability [Distributions]
• Basic Calculus [Differentiation, Integration]
• Programming
• Basic Data Structures
• Programming Knowledge
• No Prior knowledge of Computer Vision
Required

24 CS419/619: Computer Vision

 Text Books
 Digital Image Processing, R. C. Gonzalez
and R. E., Woods; Addison Wesley; 1992.
 Computer Vision: A Modern Approach, D. A.
Forsyth and J. Ponce,
 Fundamentals of Digital Image Processing,
Anil K Jain, Prentice Hall of India, 1995.

25 CS419/619: Computer Vision

 References
 Journals:
– IEEE Transactions on Pattern Analysis and Machine
Intelligence
– IEEE Transactions on Image Processing
– Pattern Recognition (PR), Elsevier
– Image and Vision Computing (IVC), Elsevier
– Pattern Recognition Letters (PRL), Elsevier
– Computer Vision, Graphics & Image Processing,
Elsevier
– International Journal of Computer Vision (IJCV),
Springer

26 CS419/619: Computer Vision

 References
 Online Resources
– CV Online:
http://homepages.inf.ed.ac.uk/rbf/CVonline/
– Computer Vision Homepage:
http://www.cs.cmu.edu/~cil/vision.html

 CS419/619 Old Videos

– https://www.youtube.com/channel/UCEgRVnL
LM6qiP6r2St8ztCg/videos

27 CS419/619: Computer Vision

 Thank you!

28 CS419/619: Computer Vision

 Tentative Assessment Policy
Credit Marks
Mid Sem 25
End Sem 40
Attendance 10
Quiz -
Assignment 10
Project Mid-sem 5
Project End-sem 10
Total 100
29 CS419/619: Computer Vision