Computer Vision (CS708)
Lecture01 - Introduction
Dr. Abdul Basit
Assistant Professor - CSIT
University of Balochistan
September 29, 2016
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 1 / 24
Introduction
Vision systems
Figure: The human visual system has no problem interpreting the subtle
variations in translucency and shading in this photograph and correctly
segmenting the object from its background..
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 2 / 24
Introduction
Vision algorithms and application.
Figure: (a) Structure from motion algorithms can reconstruct a 3D point model
of a large complex scene from hundreds of partially overlapping photographs (b)
Stereo matching algorithms can build a detailed 3D model of a building facade
from hundreds of differently exposed photographs taken from the Internet.
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 3 / 24
Introduction
Vision algorithms and application.
Figure: (c) Person tracking algorithms can track a person walking. (d) In front of
a cluttered background, Face detection algorithms, coupled with color-based
clothing and hair detection algorithms, can locate and recognize the individuals in
this image.
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 4 / 24
Contents
1 What is computer vision?
2 Applications
3 Research
4 Open research problems
5 About course
Software used
Marks distribution
Readings
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 5 / 24
Introduction
Vision systems
This kind of information we want is application specific:
3D models
Object poses
Object categories
Camera poses
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 6 / 24
Introduction
Parts of system
The “vision system” includes
Image acquisition hardware
Analog camera plus digital frame grabber, -or-
Digital camera with a fast serial interface (Firewire, USB, etc.)
Image processing support software
Your computer vision algorithms
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 7 / 24
Contents
1 What is computer vision?
2 Applications
3 Research
4 Open research problems
5 About course
Software used
Marks distribution
Readings
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 8 / 24
Introduction
Application
Important application include
Optical character recognition (OCR)
Industrial inspection and control
Mobile robot navigation
Military intelligence
Security (Surveillance)
Fingerprint recognition and biometrics
Image retrieval from digital libraries
Medical image analysis
3D model capture for visualization and animation
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 9 / 24
Introduction
Broader consumer-level applications
Important broader consumer-level applications include
Stitching
3D modeling
Face detection
Visual authentication
Morphing
Exposure bracketing
Photo based walkthroughs
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 10 / 24
Introduction
Vision systems
Figure: (a) image stitching: merging different views. (b) exposure breacketing:
merging different exposures.
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 11 / 24
Introduction
Vision systems
Figure: (c) morphing: blending between two photographs. (d) turning a collection
of photographs into a 3D model.
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 12 / 24
Contents
1 What is computer vision?
2 Applications
3 Research
4 Open research problems
5 About course
Software used
Marks distribution
Readings
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 13 / 24
Computer vision
Research
Identify an open problem in a field which is not yet addressed by
someone else and proposed a feasible solution.
To verify your solution, you need to propose a methodology to solve
the identified problem.
To support your methodology you need to conduct some experiments
and show the results.
You may need to compare your proposed method with state of the art
methods.
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 14 / 24
Computer vision
Research in computer vision
Propose an algorithm that performs better than existing algorithms.
Combine an algorithm with an existing algorithm to improve the
performance.
Combine two existing algorithms to get the desire results.
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 15 / 24
Computer vision
Research in computer vision
Borrow an algorithm form other field and implement on the vision
system to get the problem solved.
Take an existing vision algorithm and solve an existing problem.
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 16 / 24
Contents
1 What is computer vision?
2 Applications
3 Research
4 Open research problems
5 About course
Software used
Marks distribution
Readings
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 17 / 24
Introduction
Open research problems
Open research problems in computer vision
Detecting a person walking through an array of surveillance cameras.
Counting no of people (counting heads) in a crowd
Optical character recognition
Automated number plate recognition
Multiple object detection under surveillance camera.
Localizing a person using mobile camera.
Object tracking
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 18 / 24
Contents
1 What is computer vision?
2 Applications
3 Research
4 Open research problems
5 About course
Software used
Marks distribution
Readings
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 19 / 24
Contents
1 What is computer vision?
2 Applications
3 Research
4 Open research problems
5 About course
Software used
Marks distribution
Readings
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 19 / 24
Course
Software used
The software used during the course
OpenCV
Matlab or Octave
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 20 / 24
Contents
1 What is computer vision?
2 Applications
3 Research
4 Open research problems
5 About course
Software used
Marks distribution
Readings
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 20 / 24
Course
Marks distribution
Marks distribution
Mid-term (30%)
Terminal (40%)
Assignments and Quizzes (15%)
Research Report (15%)
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 21 / 24
Contents
1 What is computer vision?
2 Applications
3 Research
4 Open research problems
5 About course
Software used
Marks distribution
Readings
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 21 / 24
Introduction
Software used
Textbooks and resources used in this course
Computer Vision: Algorithms and Applications by Richard Szeliski
Introduction to Computer Vision - CS143
Computer vision from Mubarak Shah
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 22 / 24
Introduction
This semester
This semester we focus on algorithms for
3D reconstruction
Learning
Sequential state estimation (e.g. tracking)
To understand modern 3D reconstruction techniques we need to
understand how cameras transduce the world into image.
This involves understanding projective geometry and camera models.
Then we can begin to figure out how to invert what the camera does or
reconstruct the 3D scene
Dr. Abdul Basit (CSIT - UoB) Computer Vision - Lec01: Introduction September 29, 2016 23 / 24
The End
27 / 27
