Digital Image Processing ELEN E4830

Lecturers: Shahram Ebadollahi Lexing Xie

1/29/2009

1

General Information 4:10~5:20 part 1

Mondays 4:10~6:40pm Location: Mudd 1127, Room Credits: 3.0 Offered on CVN Course Webpage

5:20~5:30 break 5:30~6:40 part 2

http://www.ee.columbia.edu/~xlx/ee4830/

Target audience: First year Graduate and Senior level students

1/29/2009

2

Staff 1

Lecturers/Staff: 1

Shahram Ebadollahi Research Staff Member, IBM T.J. Watson Research PhD from Columbia U. EE dept.

1

Lexing Xie Research Staff Member, IBM T.J. Watson Research PhD from Columbia U. EE dept.

1

TA: Graham Grindlay PhD student, LabROSA

1/29/2009

3

How to reach us? 1

Shahram Ebadollahi 1 1 1

1

Lexing Xie 1 1 1

1

E-mail: [email protected] Office hours: Mondays 3:00~4:00pm Office: 1312 Mudd (EE dept., Adjunct faculty office)

E-mail: [email protected] Office hours: Mondays 3:00~4:00pm Office: 1312 Mudd

Please contact the lecturer of the week for problems/question related to each lecture!

Graham 1 1 1 1 1

1/29/2009

E-mail: [email protected] Office hours:Thursdays 2-4pm Office Location: 7LE4 CEPSR (LabROSA) Office Phone: (212) 854-0235 Mailbox:TBA

Use the discussion area in CourseWorks!

4

Our research 1

Shahram Ebadollahi 1 1

1

Image/Video content understanding Medical imaging informatics

Lexing Xie 1 1

multimedia content analysis, data mining statistical learning and signal processing in multimedia

1/29/2009

5

Pre-requisites 1 1 1 1

Signals & Systems Linear Algebra Probability If you haven’t taken these courses please see us during the break TODAY!

1/29/2009

6

Outline 1

Part I [Shahram Ebadollahi] 1 1 1 1

Course protocol, policy, and all that Introduction to DIP and examples of applications Course outline Brief review of signals and systems

1

Break

1

Part II [Lexing Xie] 1 1

1/29/2009

Introduction to MATLAB for Image Processing Brief review of linear algebra and probability 7

Protocol & Policies

1/29/2009

8

Course textbook 1

Digital Image Processing, 3rd Edition by Gonzalez and Woods, Prentice Hall 2008 (ISBN 9780131687288)

1 1

Very well written book Broad coverage of the subject Accessible by wide audience

1

Other references: see course web-page!

1

1/29/2009

9

Assignments 1

6 assignments

1

Due at: END OF DAY OF CLASS IN TA’s MAILBOX OR EMAIL INBOX! (NO EXCEPTIONS)

1

1

Solutions and graded homeworks will be handed out the week after you hand in your assignments.

Types of questions in the assignment: 1

Analytical

1

Experimental A problem which require some programming and experimentation. e.g. 1) change parameters of an image processing algorithm, observe the effect, comment on your findings 2) how do you think this image could be enhanced? Show it! Why this approach? [Programming: minimum MATLAB]

1/29/2009

10

2 Exams 1

Midterm (03/9/2009) 1 1

1

150 minutes Open book

Final (05/11/2009) 1 1

1/29/2009

3 hours Open book

11

Grading Policy 1

Homeworks: 30%

1

Midterm: 30%

1

Final: 40% 1

1/29/2009

All material will be covered in the final exam

12

DIP Introduction

1/29/2009

13

Eye •75~150 Million •Sensitive to low illumination •Distributed over Retina

•6~7 Million

• Scotopic (dim light) vision

•Highly sensitive to color •Concentrated in Fovea •Photopic (day light) vision

Rods & Cones Distribution in Retina

1/29/2009

14

Brightness

1/29/2009

15

Crab Pulsar

Electromagnetic Spectrum & more

1/29/2009

16

Gamma

X-ray

Optical

Infrared

Radio

Image 1

Image = 2d function f ( x, y ) ∈ R

x, y ∈ R

f ( x, y ) = i ( x, y ).r ( x, y ) Illumination

reflectance

0 ≤ f ( x, y ) ≤ F

1/29/2009

17

Image Digitization

1/29/2009

18

Image 1

y

Image = 2d function

f ( x, y ) ∈ R

x, y ∈ R

f ( x, y ) = i ( x, y ).r ( x, y ) Illumination

reflectance

0 ≤ f ( x, y ) ≤ F

1/29/2009

0≤ x≤ X 0≤ y ≤Y

x 19

Digital Image 1

Discretization 1 1

Spatial sampling Quantization

f (m, n) ∈ Z 0 ≤ f (m, n) ≤ L − 1

m, n ∈ Z

(0,0)

y

m

0 ≤ m ≤ M −1 0 ≤ n ≤ N −1 x

1/29/2009

Picture Element (Pixel, Pel) n

L-1

0

( M − 1, N − 1)

20

Spatial Resolution

1/29/2009

21

Spatial Resolution

1/29/2009

22

Gray-level Resolution

1/29/2009

23

Spatial and Gray-level Resolution

1/29/2009

24

Color Image B G R

255

255

255

αG αR

αB

1/29/2009

25

0

0

0

Digital Image Processing

Low level

Mid level

 x1  x  x =  2  Μ    xP 

High level 1/29/2009

26

1/29/2009

Face regions obtained using Intel’s OpenCV Library

27

Digital Image Processing

Low level

Mid level

 x1  x  x =  2  Μ    xP 

High level 1/29/2009

28

Communication & Storage

What are we going to study? Lecture No.

Date

Lecturer

Subject

1

1.26

SE/LX

Course Mechanics, Introduction to Image Processing, Introduction to MATLAB

2

2.2

SE

Digital Image Fundamentals (ch2): Sensing, Sampling, and Quantization

3

2.9

SE

Gray-level, Color and Multi-band Images (ch3 & 6), Video

4

2.16

LX Image Enhancement in Spatial Domain (ch3& 6)

1/29/2009

5

2.23

LX

Image Enhancement in Frequency Domain (ch4)

6

3.2

LX

More Image Transform, Wavelets (ch7)

Mid-term

3.9

7

3.23

LX

Image Restoration (ch5)

8

3.30

SE

Morphological Image Processing (ch9)

9

4.6

SE

Image Segmentation (ch10)

10

4.13

SE

Image Description (ch11)

11

4.20

SE

Object Recognition (ch12)

12

4.27

LX

Image Compression (ch8)

13

5.4

LX

Applications: medical, cbir, ...

29

Why study DIP? 1

Image & video is a major communication media 1

1 1

1

“An image is worth 1000 words!”

In all application domains, image and video is becoming indispensable WWW, faster computation, more storage, proliferation of image capture and consumption devices 2 Need for more, better, faster, and more intelligent image and video analysis It’s fun!

1/29/2009

30

Application Domains 1

Consumer domain 1

1

Remote sensing 1

1

1

Working methods of painters, material used, …

Security 1

1

Image mosaic, image matching, …

Art 1

1

CAD, quantification, organization, assisted surgery, …

Space explorations 1

1

Agriculture, Urban growth monitoring, …

Medical 1

1

Storage, tagging, searching,…

Surveillance, monitoring, …

Military … and many more!

1/29/2009

31

Application Domains 1

Consumer domain 1

Storage, tagging, searching,…

mountain, snow, buildings, sky

1/29/2009

32

Application Domains 1

Remote sensing 1

Agriculture, Urban growth monitoring, …

Urban growth in Washington D.C. between 1973 and 1985 Multi-spectral imaging 1/29/2009

[Image courtesy of NASA/LANDSAT] 33

Application Domains 1

Medical 1

CAD, quantification, organization, assisted surgery, …

Color doppler Echocardiogram

Tissue Microscopy 1/29/2009

Brain MRI

Spine X-ray

Cardiac CT

Brain PET (Alzheimers) 34

Application Domains 1

Space explorations 1

Image mosaic, image matching, …

Panoramic image built from images taken from mars rover Spirit [image courtesy of NASA/JPL/Cornell] 1/29/2009

35

Review of Signals & Systems

1/29/2009

36