Digital Image and Video Processing - Communications and Signal ...

Scope of ENEE631

Digital Image and Video Processing – An Introduction

z

First graduate course on image/video processing

z

Prerequisites: ENEE620 and 624, or by permission – Not assume you have much exposure on image processing at undergraduate level – Random processes and DSP are required background

z

Spring ’06 Instructor: K. J. Ray Liu

Emphasis on fundamental concepts – Provide theoretical foundations on multi-dimensional signal processing built upon pre-requisites – Coupled with assignments and projects for hands-on experience and reinforcement of the concepts

ECE Department, Univ. of Maryland, College Park

– Follow-up courses

ENEE631 Digital Image Processing (Spring'06)

image analysis, computer vision, pattern recognition multimedia communications and security


Lec1 – Introduction [2]

z

Primary – A.K. Jain: Fundamentals of Digital Image Processing, Prentice-Hall, 1989. (blue cover) – R.C. Gonzalez and R.E. Woods: Digital Image Processing, Prentice Hall, 2001. (black cover) – Y. Wang, J. Ostermann, Y-Q. Zhang: Digital Video Processing and Communications, Prentice-Hall, 2001.

z

References – A.Bovik & J.Gibson: Handbook Of Image & Video Processing, Academic Press, 2000.

z

UMCP ENEE631 Slides (created by M.Wu © 2001)

Textbooks

Introduction to Image and Video Processing

Other references – Will be announced in lectures





1

Why Do We Process Images?

A video is worth 1000 sentences?

http://marsrovers.jpl.nasa.gov/gallery/press/opportunity/20040125a.html JPL Mars’ Panorama captured by the Opportunity

z

Rich info. from visual data

z

Examples of images around us natural photographic images; artistic and engineering drawings scientific images (satellite, medical, etc.)

z



A picture is worth 1000 words.

z

Enhancement and restoration – remove artifacts and scratches from an old photo/movie – improve contrast and correct blurred images

z

Transmission and storage – images from oversea via Internet, or from a remote planet

z

Information analysis and automated recognition – providing “human vision” to machines

z

Security and rights protection – encryption and watermarking

“Motion pictures” => video movie, TV program; family video; surveillance and highway/ferry camera ENEE631 Digital Image Processing (Spring'06)



z

List of Image and Video Processing Examples

“Exactness” – Perfect reproduction without degradation – Perfect duplication of processing result

z

Convenient & powerful computer-aided processing – Can perform rather sophisticated processing through hardware or software – Even kindergartners can do it!

z

Easy storage and transmission – 1 CD can store hundreds of family photos! – Paperless transmission of high quality photos through network within seconds





Why Digital?


z

Compression

z

Manipulation and Restoration – Restoration of blurred and damaged images – Noise removal and reduction – Morphing

z

Applications – – – –

Visual mosaicing and virtual views Face detection Visible and invisible watermarking Error concealment and resilience in video transmission



2

Compression

z

Color image of 600x800 pixels UMCP ENEE631 Slides (created by M.Wu © 2001)


z

Denoising

– Without compression

600*800 * 24 bits/pixel = 11.52K bits = 1.44M bytes

– After JPEG compression (popularly used on web)

only 89K bytes compression ratio ~ 16:1

Movie – 720x480 per frame, 30 frames/sec, 24 bits/pixel – Raw video ~ 243M bits/sec – DVD ~ about 5M bits/sec – Compression ratio ~ 48:1

From X.Li http://www.ee.princeton.edu/~lixin/denoising.htm

“Library of Congress” by M.Wu (600x800)





Morphing


Deblurring


http://www.mathworks.com/access/helpdesk/help/toolbox/images/deblurr7.shtml Princeton CS426 face morphing examples http://www.cs.princeton.edu/courses/archive/fall98/cs426/assignments/morph/morph_results.html





3

Visual Mosaicing

Face Detection UMCP ENEE631 Slides (created by M.Wu © 2001)


– Stitch photos together without thread or scotch tape

Face detection in ’98 @ CMU CS, http://www.cs.cmu.edu/afs/cs/Web/People/har/faces.html

R.Radke – IEEE PRMI journal paper draft 5/01 ENEE631 Digital Image Processing (Spring'06)




Invisible Watermark UMCP ENEE631 Slides (created by M.Wu © 2001)


Visible Digital Watermarks

from IBM Watson web page “Vatican Digital Library”


– 1st & 30th Mpeg4.5Mbps frame of original, marked, and their luminance difference – human visual model for imperceptibility: protect smooth areas and sharp edges Lec1 – Introduction [15]



4

Data Hiding for Annotating Binary Line Drawings

Error Concealment

original



(a) original lenna image

pixelpixel-wise difference

marked w/ “01/01/2000” 01/01/2000”

(b) corrupted lenna image

25% blocks in a checkerboard pattern are corrupted

(c) concealed lenna image

corrupted blocks are concealed via edge-directed interpolation

Examples were generated using the source codes provided by W.Zeng.




Sampling and Quantization y

Grayscale image

x

– I(x,y) is the intensity of the image at the point (x,y) on the image plane.

I(x,y) takes non-negative values

250

assume the image is bounded by a rectangle [0,a]×[0,b]

200

I: [0, a] × [0, b] → [0, inf )


– A grayscale image is a function I(x,y) of the two spatial coordinates of the image plane.

intensity

UMCP ENEE408G Slides (created by M.Wu & R.Liu © 2002)

What is An Image? z

150 100

50 I(x,y)

z

Color image – Can be represented by three functions, R(x,y) for red, G(x,y) for green, and B(x,y) for blue. ENEE631 Digital Image Processing (Spring'06)


0 100 80

y

60

50

rows

40 20 0

0

x

z

Computer handles “digital” data.

z

Sampling – Sample the value of the image at the nodes of a regular grid on the image plane. – A pixel (picture element) at (i, j) is the image intensity value at grid point indexed by the integer coordinate (i, j).

z

Quantization

255 (white)

– Is a process of transforming a real valued sampled image to one taking only a finite number of distinct values. – Each sampled value in a 256-level grayscale image is represented by 8 bits.

0 (black)

columns




5


Examples of Quantizaion


Examples of Sampling

256x256

8 bits / pixel

64x64

4 bits / pixel 16x16


2 bits / pixel




6