digital image processing - columbia university
TRANSCRIPT
1/29/2008 1
Digital Image ProcessingELEN E4830
Professors:Shahram EbadollahiLexing Xie
1/29/2008 2
General Information
Spring 2008Mondays 4:10~6:40pmLocation: Eng. Mudd 1127Credits: 3.0Offered on CVN
4:10~5:20 part 1
5:20~5:30 break
5:30~6:40 part 2
Target audience:First year Graduate and Senior level students
Course Webpage http://www.ee.columbia.edu/~xlx/ee4830/
1/29/2008 3
Staff
� Lecturers/Staff:� Shahram Ebadollahi
Research Staff Member, IBM T.J. Watson ResearchPhD from Columbia U. EE dept.
� Lexing XieResearch Staff Member, IBM T.J. Watson ResearchPhD from Columbia U. EE dept.
� TA: Wei Liu
1/29/2008 4
How to reach us?
� Shahram Ebadollahi� E-mail: [email protected]� Office hours: Mondays 3:00~4:00pm� Office: 1312 Mudd
� Lexing Xie� E-mail: [email protected]� Office hours: Mondays 3:00~4:00pm� Office: 1312 Mudd
� Wei Liu� E-mail: [email protected]� Office hours:TBD� Office: 711 CEPSR� Mailbox:TBD
Please contact the lecturer of the week
for problems/question related to each
lecture!
1/29/2008 5
Our research
� Shahram Ebadollahi� Image/Video content understanding� Medical imaging informatics
� Lexing Xie� multimedia content analysis, data mining� statistical learning and signal processing in multimedia
1/29/2008 6
Pre-requisites
� Signals & Systems� Linear Algebra� Probability
� If you haven’t taken these courses please see us after the class TODAY!
1/29/2008 7
Outline
� Part I [Shahram Ebadollahi]
� Course protocol, policy, and all that� Introduction to DIP and examples of applications� Course outline� Brief review of signals and systems
� Break
� Part II [Lexing Xie]
� Introduction to MATLAB for Image Processing� Brief review of linear algebra and probability
1/29/2008 8
Course Protocols & Policies
1/29/2008 9
Course textbook
� Digital Image Processing, 3rd Editionby Gonzalez and Woods, Prentice Hall 2008 (ISBN 9780131687288)
� Very well written book� Broad coverage of the subject� Accessible by wide audience
� Other references: see course web-page!
1/29/2008 10
Assignments
� 6 assignments
� Due at: END OF DAY OF CLASS IN TA’s MAILBOX OR EMAIL INBOX! (NO EXCEPTIONS)� Solutions and graded homeworks will be handed out the week after you hand in your
assignments.
� Types of questions in the assignment:
� Analytical
� ExperimentalA 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?
1/29/2008 11
2 Exams
� Midterm (03/10/2008)� 150 minutes� Open book
� Final� 3 hours� Open book
1/29/2008 12
Grading Policy
� Homeworks: 30%
� Midterm: 30%
� Final: 40%� All material will be covered in the final exam
1/29/2008 13
DIP Introduction
1/29/2008 14
Eye Physiology
1/29/2008 15
Eye Physiology & Visual Perception
Rods & Cones Distribution in Retina
•6~7 Million
•Highly sensitive to color
•Concentrated in Fovea
•Photopic (day light) vision
•75~150 Million
•Sensitive to low illumination
•Distributed over Retina
• Scotopic (dim light) vision
Photoreceptor Cells
1/29/2008 16
Electromagnetic Spectrum & more
Gamma X-ray Optical Infrared Radio
Cra
b P
ulsa
r
1/29/2008 17
Image
� Image = 2d function
Ryxf ∈),( Ryx ∈,
),().,(),( yxryxiyxf =
Illumination reflectance
Fyxf ≤≤ ),(0
1/29/2008 18
Image
� Image = 2d function
Ryxf ∈),( Ryx ∈,
x
y
),().,(),( yxryxiyxf =
Illumination reflectance
Fyxf ≤≤ ),(0
Yy
Xx
≤≤≤≤
0
0
1/29/2008 19
Digital Image
x
y� Discretization
� Spatial sampling� Quantization
Znmf ∈),( Znm ∈,
)0,0(
)1,1( −− NM
m
n
10
10
−≤≤−≤≤
Nn
Mm1),(0 −≤≤ Lnmf
0
L-1
Picture Element
(Pixel, Pel)
1/29/2008 20
Digital Image
x
y
−−−−
−−
=
)1,1()1,1()0,1(
)1,1()1,1()0,1(
)1,0()1,0()0,0(
NMfMfMf
Nfff
Nfff
f
�
����
�
�
1/29/2008 21
Color Image
R
G
B
0
255
0
255
0
255
Rα
Gα
Bα
1/29/2008 22
Digital Image Processing
Low level
Mid level
=
Px
x
x
x�
2
1
High level
1/29/2008 23
1/29/2008 24
Digital Image Processing
Low level
Mid level
=
Px
x
x
x�
2
1
High level
Communication & Storage
1/29/2008 25
What are we going to study?
Applications: medical, cbir, ...SE5.513
Image Compression (ch8)LX4.2812
Object Recognition (ch12)LX4.2111
Image Description (ch11)SE4.1410
Image Segmentation (ch10)SE4.79
Morphological Image Processing (ch9)SE3.318
Image Restoration (ch5)LX3.247
3.10Mid-term
More Image Transform, Wavelets (ch7)LX3.36
Image Enhancement in Frequency Domain (ch4)LX2.255
Image Enhancement in Spatial Domain (ch3& 6)LX2.184
Gray-level, Color and Multi-band Images (ch3 & 6), Video
SE2.113
Digital Image Fundamentals (ch2): Sensing, Sampling, and Quantization
SE2.42
Course Mechanics, Introduction to Image Processing, Introduction to MATLAB
SE/LX1.281
SubjectLecturerDateLecture No.
1/29/2008 26
Why study DIP?
� Image & video is a major communication media� “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 � Need for more, better, faster, and more intelligent image and video analysis
� It’s fun!
1/29/2008 27
Application Domains� Consumer domain
� Storage, tagging, searching,…� Remote sensing
� Agriculture, Urban growth monitoring, …
� Medical� CAD, quantification, organization,
assisted surgery, …
� Space explorations � Image mosaic, image matching,
…� Art
� Working methods of painters, material used, …
� Security� Military� … and many more!
1/29/2008 28
Application Domains
� Consumer domain� Storage, tagging,
searching,…
mountain, snow, buildings, sky
1/29/2008 29
Application Domains
� Remote sensing� Agriculture, Urban growth
monitoring, …
Urban growth in Washington D.C. between 1973 and 1985
[Image courtesy of NASA/LANDSAT]Multi-spectral imaging
1/29/2008 30
Application Domains
� Medical� CAD, quantification,
organization, assisted surgery, …
Color doppler Echocardiogram Brain MRI Cardiac CT
Tissue Microscopy Spine X-rayBrain PET (Alzheimers)
1/29/2008 31
Application Domains
� Space explorations � Image mosaic, image
matching, …
Panoramic image built from images taken from mars rover Spirit
[image courtesy of NASA/JPL/Cornell]
1/29/2008 32
Review of Signals & Systems