Introduction

Prof. Iyad Jafar

OUTLINE
• Introduction
• What is Computer Vision?
• Why Study Computer Vision?
• Goals of Computer Vision
• Applications of Computer Vision
• Why is Computer Vision Hard?
• Computer Vision Research
• CV and Deep Learning
2
INTRODUCTION

3
INTRODUCTION
• Vision is the most powerful sense
• Allows interaction with and
navigating the physical world
without physical contact
• “Our sight is the most delightful
of all our senses,” Joseph Addison
• 60% of the brain involved in
visual perception

@Venti Views4
INTRODUCTION
• We can perceive the
3D structures of the
world with ease!
• You can tell a lot
with your eyes.
• Shape, depth,
structure, texture,
emotions, actions
….

@Leo Rivas 5
INTRODUCTION
• A huge effort has been done by perceptual
psychologists to understand the human vision
• However, our understanding is still not fully
understood and can be easily tricked.
• Efforts to develop mathematical techniques to
Hermann grid illusion
recover three-dimensional shape and
appearance of objects in the real world
images.
• Computer vision!
6
WHAT IS COMPUTER VISION?

7
WHAT IS COMPUTER VISION?
• Computer vision is the enterprise of building machines /
computers that can see.

• Computer vision can be viewed as

• Automating human visual processes
• Information processing task Vision
Software
• Inverting image formation
• Inverse graphics
Scene Description
• It is challenging, useful and fun! 
8
WHY STUDY COMPUTER VISION?
• The human visual system is so powerful!
• Why to build machines that can see?
• Delegate routine and daily tasks to machines
• Human visual system is qualitative rather than quantitative (it can’t
make precise measurements of the physical world)
• Machine vision can be designed to surpass the human visual system
and extract information that humans cannot see
• Billions of images/videos are generated per day; huge number of
applications
9
GOALS OF COMPUTER VISION

10
LOW-LEVEL VISION
• Measurements, enhancements, region segmentation, features
extraction ….

11
MID-LEVEL VISION
• 3D Reconstruction, depth estimation, motion estimation

12
HIGH LEVEL VISION
• Category detection, activity recognition, deep understandings

This is a building with

many windows and
grass in front of it.
There is a person
walking on the right …

13
APPLICATIONS

14
OPTICAL CHARACTER RECOGNITION

15
BIOMETRICS

16
SELF-DRIVING CARS

17
COMPUTER-AIDED DIAGNOSIS AND GUIDED SURGERY

18
SPORTS

19
AUGMENTED REALITY AND INTERACTION

20
ROBOTICS

21
OBJECT DETECTION AND SEGMENTATION

22
PANORAMAS AND 3D RECONSTRUCTION

23
TEXT TO IMAGE SYNTHESIS

Try DeepAI [Link]

24
IMAGE GENERATION AND STYLE TRANSFER

Try @ Anytools [Link]

25
WHY IS COMPUTER VISION HARD?

26
WHY IS IT HARD?

Symantec Gap 27
WHY IS IT HARD?

Viewpoint

Mapping from 3D to 2D Illumination 28

WHY IS IT HARD?

Occlusion

Intra-class variation
Scale
Deformation
29
COMPUTER VISION RESEARCH

30
COMPUTER VISION RESEARCH
• You just saw many examples of current systems
• Many of these are less than 5 years old
• Computer vision is an active research area, and is rapidly
changing
• Many new apps in the next 5 years
• Deep learning powering many modern applications
• Many startups across a dizzying array of areas
• Deep learning, robotics, autonomous vehicles, medical imaging,
construction, inspection, VR/AR
31
COMPUTER VISION RESEARCH
• 50+ years of computer vision research
• Vision is a hard problem
• Multi-disciplinary field

32
COMPUTER VISION RESEARCH

33
COMPUTER VISION RESEARCH

34
CV AND DEEP LEARNING
• Deep learning is very popular today and has been applied in
many CV applications
• Yet, why to learn the basics?
• Laborious and need to train a network with tons of data to learn a
phenomena that can be precisely described by the basics
• When a network does not perform well, the basics may help
• Collecting and annotating data could be expensive; use the basics to
synthesize data
• Curiosity!
35
READINGS
• SZ
• 1.1, 1.2
• The Computer Vision Industry (
https://www.cs.ubc.ca/~lowe/vision.html)

36
VIDEOS
• What is computer vision? (https://youtu.be/wVE8SFMSBJ0 )
• What is vision used for? (https://youtu.be/qt1UfF0fn4w)

37