Computer Vision
Lecture 1. Introduction
Hale Kim, hikim@inha.ac.kr
Xuenan Cui, xncui@inha.ac.kr
and
Edison Li, szli@vision.inha.ac.kr
Computer Vision Lab.
Dept of Information and Communication Eng.
INHA University
Why study Computer Vision?
• Images and movies are everywhere
• Fast-growing collection of useful applications
–
–
–
–
building representations of the 3D world from pictures
automated surveillance (who’s doing what)
movie post-processing
face finding
• Various deep and attractive scientific mysteries
– how does object recognition work?
• Greater understanding of human vision
Computer Vision - A Modern Approach
Set: Introduction to Vision
Slides by D.A. Forsyth
Properties of Vision
• One can “see the future”
– Cricketers avoid being hit in the head
• There’s a reflex - when the right eye sees something
going left, and the left eye sees something going
right, move your head fast.
– Gannets pull their wings back at the last
moment
• Gannets are diving birds; they must steer with their
wings, but wings break unless pulled back at the
moment of contact.
• Area of target over rate of change of area gives time
to contact.
2022-12-22
3
Properties of Vision
• 3D representations are easily constructed
– There are many different cues.
– Useful
• to humans (avoid bumping into things; planning
a grasp; etc.)
• in computer vision (build models for movies).
– Cues include
• multiple views (motion, stereopsis)
• texture
• shading
Computer Vision - A Modern Approach
Set: Introduction to Vision
Slides by D.A. Forsyth
Properties of Vision
• People draw distinctions between what is seen
–
–
–
–
–
–
–
“Object recognition”
This could mean “is this a fish or a bicycle?”
It could mean “is this George Washington?”
It could mean “is this poisonous or not?”
It could mean “is this slippery or not?”
It could mean “will this support my weight?”
Great mystery
• How to build programs that can draw useful distinctions
based on image properties.
Computer Vision - A Modern Approach
Set: Introduction to Vision
Slides by D.A. Forsyth
Computer Vision
One picture is worth more than millions
of words!
6
Computer Vision
• Primary Objectives
– Improvement of pictorial information for human
interpretation.
• Noise removal
• Night vision
– Processing of scene data for storage, transmission,
and representation for autonomous machine
perception.
• Automated process
• Robot vision, Face recognition
7
Computer Vision
• Examples
Adaptive HPF
Multi-beam echo-sounder image of seafloor
8
What is digital image processing?
•
f(x,y): An image defined as a 2D function
x, y: spatial coordinates
f: intensity or gray level of the image at (x,y)
•
•
Digital image if x, y, and f are all discrete
Pixel (picture cell or picture element): unit of digital image
Sampling
Scene
Sensor
1
2

511
255
255
255

255
1
255
188
124

255
2
255
60
0

255


511
255
x y
0
0
Quantization
Image
file



255
255
255

255
Digital image in a matrix
9
Origin of Computer Vision
• History
– Early 20’s : beginning of DIP
• Ex: Bartlane cable pictures transmission system (1921): invented by Mr.
Bartholomew and Captain MacFarlane of the Daily Mirror of London, England.
• Submarine cable between London and Halifax, N.S., Canada.
10
http://www.hffax.de/history/html/bartlane.html
Origin of Computer Vision
• History
– Early 20’s : beginning of DIP
• Baudot tape from telegraphic typewriter  Photographic
reproduction (5 levels)  15 levels in 1929
1921
11
1922
1929
Origin of Computer Vision
• History
– 60’s: Development of Digital computer + Space programs lead
by NASA and Jet Propulsion Lab.
• Ex: Ranger 7 (1964) - transmitted the picture of Moon, Mariner
series (Mars), Apollo, Pioneers (Jupiter, Saturn), Viking (Mars),
Voyagers (Uranus, Neptune).
12
First picture of Moon
by Ranger 7 (1964)
First picture of Mars
by Mariner 4 (1965)
Origin of Computer Vision
• History
– 60’s: Development of Digital computer + Space programs lead
by NASA and Jet Propulsion Lab.
13
Ariel: A moon of Uranus by Voyager (2000)
Neptune by Voyager2 (1999)
Origin of Computer Vision
• History
– 70’s and 80’s: progress in algorithms and hardware, and
tremendous applications to various fields.
• Ex: Robot vision, Factory automation, Military reconnaissance,
Medical imaging (CT, X-ray, Angiography, ...), Agriculture (Crop
assessment), Ecology (Whale migration), etc.
14
Origin of Computer Vision
• History
– 70’s and 80’s: progress in algorithms and hardware, and
tremendous applications to various fields.
• Ex: Robot vision, Factory automation, Military reconnaissance,
Medical imaging (CT, X-ray, Angiography, ...), Agriculture (Crop
assessment), Ecology (Whale migration), etc.
15
First LANDSAT7 image (1999)
Automated Optical Inspection
Origin of Computer Vision
• History
– 90’s: Real-time imaging systems.
• Ex: Motion pictures, Virtual reality, Digital medical imaging
systems, 3D vision, etc.
16
Machine vision for Traffic
surveillance
Autonomous mobile robot
Examples of Field using
Computer Vision
• Military
– reconnaissance (Satellite imagery), target tracking (smart bomb,
cruise missile)
• Remote sensing
– Land-cover analysis, terrain rendering, GIS, OIS, weather
• Forensics/Security systems
– fingerprint recognition, eye recognition, snapshots of moving cars’
license plates
• Medical diagnostic imaging
– X-ray, Ultrasound, Computed tomography, Nuclear Magnetic
Resonance, Angiogram
• Biological research
– cell analysis, DNA classification and matching
• Factory automation
– automatic inspection, part-assembly and CAD model-based
inspection
17
Examples of Field using
Computer Vision– cont.
• Navigation
– autonomous mobile robots, unmanned vehicles
• Ecology
– study of animal migration
• Document imaging and iDBMS
– archiving and retrieval, character recognition
• Photography/Motion pictures
– digital camera, image synthesis, morphing and warping
• Teleperception
– environment restoration, waste management, automatic event
detection in ocean and space
– Any new applications of DIP these days?
18
Examples – Japan Earthquake in March 11, 2011
19
Before
After
20
Examples of Field using
Computer Vision– cont.
21
Examples of Field using
Computer Vision– cont.
22
Examples of Field using
Computer Vision– cont.
• Gamma-ray
Imaging
– Nuclear medical
imaging
PET
Positron Emission Tomography
– Astronomical
observations
– Gamma radiation
from a reactor valve
23
Examples of Field using
Computer Vision– cont.
• X-ray Imaging
– Chest X-ray
– Angiography
– Computerized Axial
Tomography (CAT)
– AXI
– Astronomical observations
24
Examples of Field using
Computer Vision– cont.
• Imaging in the Ultraviolet
Band
–
–
–
–
–
–
25
Lithography
Microscopy
Lasers
Biological imaging
Industrial inspection
Astronomical observations
Examples of Field using
Computer Vision– cont.
• Imaging in the Visible and
Infrared Bands
– Most popular usage
– Combined usage between
visible and infrared
26
Examples of Field using
Computer Vision– cont.
 Imaging in the Visible and
Infrared Bands
Remote sensing
Multi-spectral image
27
Examples of Field using
Computer Vision– cont.
 Imaging in the Visible and
Infrared Bands
Weather observations
28
Examples of Field using
Computer Vision– cont.
 Imaging in the Infrared
Bands
Estimation of regional
population and energy
consumption
29
Examples of Field using
Computer Vision– cont.
 Imaging in the Infrared
Bands
Estimation of regional
population and energy
consumption
30
Examples of Field using
Computer Vision– cont.
 Imaging in the
Visible and
Infrared Bands
Manufacturing
automation
31
Examples of Field using
Computer Vision– cont.
 Imaging in the Visible and
Infrared Bands
Forensics
License plate reading
32
Examples of Field using
Computer Vision– cont.
 Imaging in Terra Hertz (1012)
Airport security
33
Examples of Field using
Computer Vision– cont.
• Imaging in the Microwave Band
– Radar
34
Examples of Field using
Computer Vision– cont.
• Imaging in the Radio Band
– Medicine: Magnetic
Resonance Imaging(MRI)
– Astronomy
35
Examples of Field using
Computer Vision– cont.
• Examples in which other imaging modalities are used
Ultrasound
CG
36
Examples of Field using
Computer Vision– cont.
• Examples in which other imaging modalities are used
37
Fundamental Steps in CV
Domain-specific
or
Problem-oriented
38
Fundamental Steps in CV
•
Image Digitization
– To convert continuous brightness and spatial coordinates into discrete
components
•
Sampling : spatial discretization
– High sampling rate (Over-sampling) - requires extra time & memory space
– Low sampling rate (Under-sampling) - causes aliasing (distortion)
•
Quantization : converts continuous light intensity to discrete gray levels
Sampling
x y
0
1
2

511
0
255
255
255

255
1
255
188
124

255
2
255
60
0

255






511
255
255
255
255
255
Quantization
Digital image in a matrix
39
Fundamental Steps in CV
• Image Enhancement
– To enhance certain image features for subsequent analysis or for
image display.
– NOT increases the inherent information content in the image.
– Ex: Contrast stretching, Edge enhancement, Smoothing, Sharpening, ...
40
Fundamental Steps in CV
• Image Restoration
– To improve the quality of an image by removing or minimizing
known degradations (by limitations of a sensor or geometric
distortion) in the image.
– Models the inverse process of a priori known degradation, and
reverses the effect of the degradation.
41
Fundamental Steps in CV
•
Image Coding
– To reduce the number of bits in a digital image for the purpose of minimizing
storage or channel capacity in transmission.
– Lossless: preserve the exact data in the original image.
– Lossy: the quality of the decoded image may not be precisely identical to
that of the original image.
– Ex: Karhunen-Loeve Transform, Discrete Fourier Transform, Discrete Cosine
Transform, ...
42
Fundamental Steps in CV
• Image Reconstruction
– To reconstruct 2D or 3D objects from several 1D projections
– Ex: CT, MRI, PET, Scientific visualization and Animation.
43
Fundamental Steps in CV
• Image Segmentation
– To subdivide an image into its constituent regions or objects.
44
Fundamental Steps in CV
• Image Representation & Description
– To symbolically represent the contents of an image for highlevel computer processing.
45
Fundamental Steps in CV
• Pattern Recognition
– To enable machines (or computers) to automatically recognize
and understand scenes.
– Pattern matching, Description, Feature extraction, Decisionmaking.
– Ex: Character recognition, Fingerprint recognition, ...
• Statistical, Syntactic, Optical, Neural Networks
46
Components of CV System
47
Components of CV System
• Image acquisition equipment: lights, sensor, frame grabber
– Digitizer : spatially samples an image into pixels and converts
light intensity to electrical signal.
– Ex: Scanner, Vidicon camera, CCD-camera.
– Frame Grabber : A/D converter + Frame buffer + D/A converter
• Processing equipment: Software and DSP board
• Output or Display equipment : Process controller, Alarm,
CIM system, Monitor, Printer
• Storage devices :
– Image of 512x512 pixels (1 byte/pixel) = 262,144B (0.25 MB)
– 5.5 frames/1.44MB, 2800 frames/700MB = 93.3 secs of movie.
48
Any CV applications from your daily lives?
49