Computer Vision I Introduction

peachpuceΤεχνίτη Νοημοσύνη και Ρομποτική

6 Νοε 2013 (πριν από 3 χρόνια και 10 μήνες)

80 εμφανίσεις

Computer Vision I

Introduction

Raul Queiroz Feitosa

11/6/2013

Introduction

2

Content


What is CV?


CV Applications


Fundamental Steps



From DIP to CV



Course Content

11/6/2013

Introduction

3

What is Computer Vision



“Computer Vision is the science that develops
the theoretical and algorithmic basis by which
useful information about the world can be
automatically extracted and analyzed from an
observed image, image set, or image sequence
from computations made by a ... computer
.”
R. B.
Haralick, L.G. Shapiro


11/6/2013

Introduction

4

Applications


Medical Image Analysis


Analysis of Remote Sensing Data


Biometrics


Security


Microscopy


Industrial Inspection




11/6/2013

Introduction

5

Applications

Medical Images

Microscopy


Industry

Security

Robot Vision

Biometrics

Remote Sensing

much
more


11/6/2013

Introduction

6

LVC Topics:
Face Recognition

11/6/2013

Introduction

7

Controle de
Passaportes

Registro Único de Identidade Civil
RIC

Controle de Acesso

Aplicações Criminais

LVC Topics:
Face Recognition

11/6/2013

Introduction

8

Suspect Behavior

Tracking

Recognition

Frontal View

LVC Topics:
Face Recognition from
Video

11/6/2013

Introduction

9

LVC Topics:
Medical Image Analysis

LVC Topics:
Remote Sensing

11/6/2013

Introduction

10

11/6/2013

Introduction

11

LVC
Applications:

Remote Sensing

Geometric features are used to
distinguish landing lanes from other
targets in the forest.


Illegal runways

SAR R99B (SIPAM)

Alves et al., 2009

11/6/2013

Introduction

12

Fundamental Steps

Image Acquisition: digitizes the electromagnetic
energy


(quem /
o que)

Physical image

digital image

gray level

physical
image

digital
image

(pixels)


Acquisition

Enhancement

Segmentation

Feature
extraction

Recognition

Post
-

processing

11/6/2013

Introduction

13

Fundamental Steps

Image Enhancement: improves image quality


digital
image



digital
image

Acquisition

Enhancement

Segmentation

Feature
extraction

Recognition

Post
-

processing

11/6/2013

Introduction

14

Fundamental Steps


Segmentation: partitions the image into
meaningfull objects

segments

digital image

Acquisition

Enhancement

Segmentation

Feature
extraction

Recognition

Post
-

processing

11/6/2013

Introduction

15

Fundamental Steps

Post
-
Processing: support segmentation/description

segments

segments

Acquisition

Enhancement

Segmentation

Feature
extraction

Recognition

Post
-

processing

11/6/2013

Introduction

16

Fundamental Steps

Description: converts the data into a form suitable
for processing

segments

description

Acquisition

Enhancement

Segmentation

Feature
extraction

Recognition

Post
-

processing

x
1
=
(
x
11
… x
1n
)
T

x
i
=
(
x
i1
… x
in
)
T

x
p
=
(
x
p1
… x
pn
)
T


∙ ∙


∙ ∙

11/6/2013

Introduction

17

Fundamental
Steps

Recognition: assigns a label to the image objects

description

label

Acquisition

Enhancement

Segmentation

Feature
extraction

Recognition

Post
-

processing

x
1
=
(
x
11
… x
1n
)
T

x
i
=
(
x
i1
… x
in
)
T

x
p
=
(
x
p1
… x
pn
)
T


∙ ∙


∙ ∙

paprika


pepper


cabbage


∙ ∙


∙ ∙

11/6/2013

Introduction

18

From DIP to CV

Digital Image Processing


Input and output are images!


From image up to recognition!

Acquisition

Enhancement

Segmentation

Feature
extraction

Recognition

Post
-

processing

DIP

DIP

11/6/2013

Introduction

19

From DIP to CV

Image Analysis/Understanding


From segmentation up to recognition.

Acquisition

Enhancement

Segmentation

Feature
extraction

Recognition

Post
-

processing

Image Analysis

11/6/2013

Introduction

20

From DIP to CV

Computer Vision


Tries to emulate human intelligence.


Emphasis on 3D analysis.

Acquisition

Enhancement

Segmentation

Feature
extraction

Recognition

Post
-

processing

Computer Vision

11/6/2013

Introduction

21

From DIP to CV

Process Levels


Low
-
level: input and outputs are images


Mid
-
level: image as input and attributes as output
.


High
-
level: “making sense” of an ensemble of objects

Acquisition

Enhancement

Segmentation

Feature
extraction

Recognition

Post
-

processing

Low




Mid




High


11/6/2013

Introduction

22

Image Analysis



develops methods and algorithms able to extract
automatically useful information about the world
.

Image
Analysis

11/6/2013

Introduction

23

Computer Graphics



develps techniques for visualization and manipulation
of ideas that exist only conceptually or in form of
mathematical description, but not as concrete object.

Computer



Graphics

11/6/2013

Introduction

24

Course Content

Main:


Introduction


Digital Image Fundamentals


Image Enhancement in Spatial Domain


Image Enhancement in Frequency Domain


Morphological Image Processing


Segmentation


Representation and Description


Object Recognition

Appendices:


Mathematical Foundation


Dimensionality Reduction
(
top
)

11/6/2013

Introduction

25

Bibliography

1.
R. G. Gonzalez, R. E. Woods,
Digital Image Processing;
Prentice Hall, 3rd Ed.,
2007

2.
R. G. Gonzalez, R. E. Woods,
Digital Image Processing;
Prentice Hall, 2nd Ed.,
2002.

3.
R. G. Gonzalez, R. E. Woods, S.L. Eddings,

Digital Image Processing using
MATLAB;
Prentice Hall, 2003.

4.
M. Nixon, A. Aguado,
Feature Extraction & Image Processing
, Newnes, 2002.

5.
R. O. Duda, Peter E. Hart, D. G. Stork,

Pattern Classification
, Wiley
-
Interscience; 2nd edition, 2000.

11/6/2013

Introduction

26

Next Topic

Digital

Image

Fundamentals