Sie sind auf Seite 1von 9

EE4206 Digital Image Processing

Course Intended Learning Outcomes (CILOs)



Number CILOs
1 Describe principles of different digital imaging systems
2 Analyse and design digital filters for two-dimensional (2-D) signals
3 Perform the 2-D discrete Fourier transform (DFT)
4 Implement image processing algorithms on computers
5 Apply computer algorithms to practical problems

Teaching and Learning Activities (TLAs)

CILOs 1, 2, 3 Lectures, tutorials, on-line learning
CILOs 4, 5 Computer projects

Assessment Tasks/Activities

Type of assessment tasks Weighting
Continuous assessment Written assignments, quizzes, computer projects 30%
Examination Written examination 70%

Grading of Student Achievement

Letter Grade Grade Point Grade Definitions
A+
A
A-
4.3
4.0
3.7
Excellent
B+
B
B-
3.3
3.0
2.7
Good
C+
C
C-
2.3
2.0
1.7
Adequate
D 1.0 Marginal
F 0.0 Failure
EE4206 Digital Image Processing
Tutorial #1



Assume that a video system has
- Each frame size 512 by 512;
- 30 frames/second;
- 8 bits for each color component.

(1) What is the minimum data rate needed to transmit the
video if the data are uncompressed?
(2) What compression ratio is needed to transmit the data
through a modem with the speed of 56K bit/s?
(3) Suppose that we can down-sample the image to reduce the
amount of data. For a compression ratio of 100, what is
the maximum image size the above modem can handle?
EE4206 Digital Image Processing

Tutorial #2


1. Estimate the diameter of the smallest dot on a page that the eyes can discern if the
page is 0.2m away from the eyes. Assume:
- The retina has a diameter of 28 mm.
- The fovea is a square area of 1.5mm by 1.5 mm.
- The fovea has 580 by 580 receptors equally spaced in the square area. The
eyes cannot resolve a dot whose image on the fovea is less than a receptor in
size.

2. Suppose that a flat area with center at (x
0
, y
0
) is illuminated by a light source with
intensity distribution
] ) ( ) [(
2
0
2
0
) , (
y y x x
Ke y x i
+
=

Assume that the reflectance of the area is 1, and let K = 255. If the resulting
image is digitized with m bits of intensity resolution, and the eye can detect an
abrupt change of eight shades of intensity between adjacent pixels, what value of
m will cause visible false contours?



Left image: m is too small (too few gray levels); right image: m is enough.

EE4206 Digital Image Processing

Tutorial #3


1. Suppose that a flat area with center at (x
0
, y
0
) is illuminated by a light source with
intensity distribution
] ) ( ) [(
2
0
2
0
) , (
y y x x
Ke y x i
+
=

Assume that the reflectance of the area is 1, and let K = 255. Assume that the
resulting image is digitized with m = 2 bits of intensity resolution with a uniform
quantiser. Determine the decision and output levels of the quantizer and sketch
the quantized image.


2. Red and green lights can be considered as radio waves with two different
frequencies. Explain why we get yellow if we mix red and green.
(Hint: A radio wave can be modeled using a sine function sin(2tvx), where v is
the frequency of the wave).
EE4206 Digital Image Processing

Tutorial #4


1. Convert the following 4-path to an 8-path.















2. Design a method to determine whether each connected component in the diagram
below
a. touches the boundary of the image;
b. contains one or more holes.





Image
boundary
EE4206 Digital Image Processing

Tutorial #5


1. Propose a set of gray-level slicing transformations capable of producing all the
individual bit planes of an 8-bit monochrome image.

Hint: the transformation for bit 7 is shown below:





2. Consider all bit planes in an image

(a) How would the histogram of the image change if we set the least significant
bit of each pixel to zero?

(b) How would the histogram of the image change if we set the most significant
bit of each pixel to zero?







1
0
128 255
EE4206 Digital Image Processing

Tutorial #6

1. In local image enhancement equalization, we need to compute the histogram of a


small sliding window in an image. Propose a method to efficiently update the
local histogram.


2. An image with intensities in the range [0, 1] has the gray scale level PDF p
r
(r)
shown in the following diagram. It is desired to transform the gray levels of this
image so that they will have the specified p
z
(z) shown. Assume that r and z are
continuous variables. Find the transformation in terms of r and z to accomplish
this.
0
1
r
p
r
(r)
2
0
1
z
p
z
(z)
2
EE4206 Digital Image Processing

Tutorial #7

1. Show that the Laplacian operator


2
2
2
2
2
y
f
x
f
f
c
c
+
c
c
= V is isotropic. That is, if we
rotate the coordinate system (x, y) by an angle u, then the Laplacian for the rotated
coordinate system is the same as that before the rotation.

Hint: Coordinate rotation can be described by


u u
u u
cos sin
sin cos
y x y
y x x
' + ' =
' ' =



2. Consider unsharp masking and highboost filtering

( ) ( ) ( ) ( )
( ) ( ) ( ) y x g k y x f y x g
y x f y x f y x f y x g
, , ,
image blurred , , , ,
mask
mask
+ =
= =


Design a 3 3 filter mask to perform the above operation.
EE4206 Digital Image Processing

Tutorial #8



1. A real function f(x) can be decomposed as the sum of an even function and an odd
function.
(a) Show that
| | | | ) ( ) (
2
1
) ( and ) ( ) (
2
1
) ( x f x f x f x f x f x f
odd even
= + =
(b) Show that
| | | | { } | | | | { } ) ( Im ) ( and ) ( Re ) ( x f F j x f F x f F x f F
odd even
= =

2. Assume that x and y are continuous variables, that | | ) ( ) ( e F x f F = and that
| | ) , ( ) , ( v u F y x f F = . Determine the Fourier transforms of
(a) ) (x f
dx
d

(b) ) , ( ) , ( y x f
y
y x f
x c
c
+
c
c

Das könnte Ihnen auch gefallen