Digital Image Processing, 2nd ed. www.imageprocessingbook.

com

Digital Image Processing

ICT-4205

(Some slides are from Professor Ochem chae,

Image Processing lab, Kyung Hee Univ.)

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Application Areas

- Military Application
- Target detection and tracking
- Surveillance
- CCTV
- Moving object detection
- Analysis of suspicious events
- Smart Car
- Medical Image Processing
- Industrial application
- Inspection
- Robot guidance
- Data Analysis
- Biometric verification and emotion analysis
© 2002 R. C. Gonzalez & R. E. Woods
 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Chapter 1: Introduction

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Image Enhancement

• Enhancement: Improve the visual quality of the image

• Eg. Noise removal using median filtering
• (from http://www.nist.gov/lispix/imlab/noise/shotfc.html)

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Image Restoration

• same as image enhancement, but you have

additional information concerning the quality
degradation. Example: removing motion
blur in an image of a fast moving object.

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Image Processing (cont.)

• Reconstruction: Reconstruction from projections.

Used in constructing 3D data from 2D projections in
computer tomography.
• Image representation using features:
– Low level representations using color, texture,
shape, motion, etc.
– High level features for recognitions; e.g., facial
features.
• Recognition and scene understanding

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Signal

ü A signal is a function that carries information.

ü Usually content of the signal changes over some set of
spatiotemporal dimensions.
üTime-Varying Signal: f(t)
Ex) audio signal

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Spatially-Varying Signal

Ø Signals can vary over space as well.

Ø An image can be thought of as being a function of 2
spatial dimensions:
f(x,y)
Ø for monochromatic images, the value of the function is
the amount of light at that point.
Ø medical CT and MRI scanners produce images that are
functions of 3 spatial dimensions:
f(x,y,z)
Ø Spatiotemporal Signals:
f(x,y,t)
ex) a video signal, animation
© 2002 R. C. Gonzalez & R. E. Woods
 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Analog and Digital Signal

Ø Most naturally-occurring signals also have a real-

valued range in which values occur with infinite
precision.
ØTo store and manipulate signals by computer we
need to store these numbers with finite precision.
thus, these signals have a discrete range.

signal has continuous domain and range = analog

signal has discrete domain and range = digital

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Sampling

Ø sampling = the spacing of discrete values in the domain

of a signal.
Ø sampling-rate = how many samples are taken per unit of
each dimension. e.g., samples per second, frames per second

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Quantization

Ø Quantization = spacing of discrete values in the

range of a signal
Ø Usually thought of as the number of bits per
sample of the signal
e.g., 1, 8, 24 bit images, 16-bit audio.

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Digital Image Representation

Ø A digital image is an image f(x,y)

that has been digitized both in
spatial coordinates and brightness.
Ø The value of f at any point (x,y) is
proportional to the brightness (or
gray level) of the image at that point.

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Components of Image Processing System

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

2D Image Sensor

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Coordinate Conversion used in this book

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Digital Image Representation

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Digital Image Acquisition Process

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Sampling and Quantization

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Example of Digital Image

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Sampling Effect

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Sampling Effect

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Quantization Effect

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Quantization Effect

Ø if the gray scale is not

enough, the smooth area will
be affected.
Ø False contouring can occur
on the smooth area which has
fine gray scales.
© 2002 R. C. Gonzalez & R. E. Woods
 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Light-intensity function

Ø image refers to a 2D light-intensity

function, f(x, y)
Ø the amplitude of f at spatial coordinates
(x, y) gives the intensity (brightness) of the
image at that point.
Ø light is a form of energy thus f(x, y) must
be nonzero and finite.
0 < f(x, y) < ∞

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Illumination and Reflectance

Ø The basic nature of f(x,y) may be

characterized by 2 components:
ü the amount of source light incident
on the scene being viewed
à Illumination, i(x,y)
ü the amount of light reflected by the
objects in the scene
à Reflectance, r(x,y)
© 2002 R. C. Gonzalez & R. E. Woods
 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Illumination and Reflectance

f(x,y) = i(x,y)r(x,y)
Ø i(x,y):
üdetermined by the nature of the light source
ü bounded by
0 < i(x,y) < ∞
Ø r(x,y) :
üdetermined by the nature of the objects
übounded by
0 < r(x,y) < 1

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Gray Level

Ø we call the intensity of a monochrome image f at

coordinate (x,y) the gray level (l) of the image at
that point.
Ø thus, l lies in the range Lmin ≤ l ≤ Lmax
Ø Lmin and Lmax are positive and finite.
Ø gray scale = [Lmin, Lmax]
Ø common practice, shift the interval to [0, L]
Ø 0 = black , L = white

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Resolution

Ø Resolution (how much you can see the detail of

the image) depends on sampling and gray levels.

Ø the bigger the sampling rate (n) and the gray scale
(g), the better the approximation of the digitized
image from the original. But the size of the image
gets bigger

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Basic Relationship among pixels

Ø Neighbors of a pixel
Ø Connectivity
Ø Labeling of Connected Components
ØDistance Measures
Ø Arithmetic/Logic Operations

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Neighbors of a Pixels

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Connectivity

Ø Let V be the set of gray-level values used to defined connectivity

ü 4-connectivity :
§ 2 pixels p and q with values from V are 4-connected if q is in the
set N4(p)
Ø8-connectivity :
ü 2 pixels p and q with values from V are 8-connected if q is in the set
N8(p)
Ø m-connectivity (mixed connectivity):
ü 2 pixels p and q with values from V are m-connected if
§ q is in the set N4(p) or
§ q is in the set ND(p) and the set N4(p)∩N4(q) is empty.
§ (the set of pixels that are 4-neighbors of both p and q whose
values are from V )

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Example of Connectivity

Path 8 neighbors m neighbors

Ø m-connectivity eliminates the multiple path

connections that arise in 8-connectivity.

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Path

Ø a path from pixel p with coordinates (x,y) to pixel

q with coordinates (s,t) is a sequence of distinct
pixels with coordinates
(x0,y0),(x1,y1),…(xn,yn)
, where (x0,y0) = (x,y) , (xn,yn) = (s,t) and (xi,yi) is
adjacent to (xi-1,yi-1)
Ø n is the length of the path
Ø we can define 4-,8-, or m-paths depending on
type of adjacency specified.

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Adjacent

Ø A pixel p is adjacent to a pixel q if they are

connected.
ØTwo image area subsets S1 and S2 are
adjacent if some pixel in S1 is adjacent to
some pixel S2.

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Distance Measures (1/3)

Ø for pixel p, q and z with coordinates (x,y),

(s,t) and (u,v) respectively,

Ø D is a distance function or metric if

ü (a) D(p,q) ≥ 0 ; D(p,q) = 0 iff p=q
ü (b) D(p,q) = D(q,p)
ü (c) D(p,z) ≤ D(p,q) + D(q,z)

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Distance Measures (2/3)

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Distance Measures (3/3)

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

M-connectivity distances

Ø distances of m-connectivity of the path

between 2 pixels depends on values of pixels
along the path.
Ø e.g., if only connectivity of pixels valued 1 is
allowed. find the m-distance between p and p4
0 p 3 p4 0 0 1 0 1 1 0 1 1
p1 p2 0 0 1 0 0 1 0 1 1 0
p 0 0 1 0 0 1 0 0 1 0 0
d=2 d=3 d=4
© 2002 R. C. Gonzalez & R. E. Woods
 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Regions
§ Let R be a subset of pixels in an image
§ We call R a region of the image if R is a connected set

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Boundary

• The boundary (also called the border or contour) of a regon R is

the set of points that are adjacent to points in the complement of R

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Region & Boundary

Regions
Boundary

Fig. Region based segmentation

© 2002 R. C. Gonzalez & R. E. Woods

 Digital Image Processing, 2nd ed. www.imageprocessingbook.com

Linear and Nonlinear Operations

Ø An operator H is said to be a linear operator

if it satisfy the following conditions.
H(af + bg) = aH(f) + bH(g)
üwhere f and g are images and a and b are scalars
üThe sum of two images is identical to applying
the operator to the image individually,
multiplying the results by the constants, and adding
those results.
Ø An operator that fails the test of above
equation is nonlinear.

© 2002 R. C. Gonzalez & R. E. Woods