Image Digitization
DR. SAKSHI
ARORA
Image Formation
Sampling and Quantization
Sampling and Quantization
Whittaker-Shanon Sampling
Frequency ≥ (2 X W)
≤ 2w
1
T
2-D sampling f(x,y) Band limited function…….F(u,v) = 0
zczc
Nyquist Theorm
• Δ�� ≤ 1/2�� ������ Δ�� ≤ 1/2��
• Are the lower bounds of the equation and are known as the Nyquist rate.
Tessellation
The spatial arrangement of the sample points is k/a Tessellation.
Calculation time!
Q1. An image is 2400 pixels wide and 2400 pixels high. The image is 300 dpi. What is the physical
size of the image?
��ℎ������������ �������� =������������ ���� ������������ ���� ��������ℎ
Sol:
������������ ���� ������������ ���� ℎ������ℎ��
�������������������� ��
��������������������
2000
=
2400
300��
300= 8 ������ℎ �� 8 ������ℎ
Hence, the physical size of the image is 8 X 8 inch2
Image Digitization
Resampling
During the scaling of an image, the number of pixels should be
increased sometimes to retain image quality: DOWNSAMPLING
Sometimes, the pixels are reduced for better compression :
UPSAMPLING
3333
�� =
339999
999933
F’ =
3+3+9+9
4
3+3+9+9 4
3+3+9+9
4
3+3+9+9 4
=6666
Resampling ( UpSampling)
Zero Order Hold Process
F =21
13
2010
F’ = F’’ =
000010300000
2211221111331133
Resampling
(Upsampling) Linear
Interpolation:
H =21
13
2010
H’ = H’’ = H’’ =
0000
1030
0000
2 1.5 1 0.5
0000
1 2 3 1.5
0000
2 1.5 1 0.5 1.5 1.75 2 1 1 2 3 1.5
0.5 1 1.5 0.75
Q2. Given a gray scale image on paper whose physical dimension is
2.5 inch X 2 inch, scanned at the rate of 150 dpi. Calculate the following:
How many bits are required to represent the image?
How much time is required to transmit the image if the modem is 28kbps?
Estimate these two values if it were a binary image.
Solution:
(a) The number of bits required to represent a gray scale image (A pixels in a
grayscale image uses 8 bits) is given as follows:
2.5 �� 150 �� 2 �� 150 �� 8 = 375 �� 300 �� 8 = 112,500 �� 8 �������� = 9,00,000
��������
(b) The total time taken to transmit the image is as follows:
���������� ������������ ���� �������� ����
����������
������������������������ ����������=9,00,000
28,000= 32.14 ��
(c)If it were a binary image, then the number of bits
required to represent the image can be calculated
as follows:
2.5 �� 150 �� 2 �� 150 �� 1 = 375 �� 300 �� 1 = 112,500 ��������
Thank You
