PSG COLLEGE OF TECHNOLOGY

DEPARTMENT OF INSTRUMENTATION AND CONTROL

SYSTEMS ENGINEERING
19U504 - DIGITAL SIGNAL PROCESSING
PROPERTIES OF 2D SYSTEMS : LINEARITY

AKALYA.M 22U201
SOWNDARIYA SREE. D 22U244
1 Date : 26.09.2024 THENDRAL TEJASREE.R 22U255

1
 AGENDA
• What is 2-D system?

• Properties of 2-D systems

• Superposition principle

• Homogeneity

• Additivity

• Problem

2 2
2-D Signal :
A 2D signal is a signal defined by two independent
variables, typically representing spatial coordinates (x, y).

They are processed using techniques like 2D Fourier

transforms, filtering, and convolution to extract information or
perform operations like enhancement, compression, or
analysis.

The formula for a 2D signal is:

f(n1, n2)

3 3
Fig 1 : 2D signals
Why we require 2D ?
2D signal processing is essential for handling signals
that vary in two dimensions, such as images and videos.
It allows for operations like filtering, compression, and
enhancement, which are crucial for tasks like image
recognition, medical imaging, and video analysis.

4 Fig 2 : 2D systems 4
2-D System :
2D systems in digital signal processing (DSP) are systems
that operate on two-dimensional data, typically images or
video frames.
These systems perform operations on the spatial
dimensions of the data, such as filtering, enhancement, and
compression.
A system T that maps an input x(n1,n2) to an output
y(n1,n2) is represented by,
y(n1,n2) = T[x(n1,n2)]

5 fig 3 : 2D system
5
 Properties of 2D System:
● Linearity

● Time invariance

● Stability

● Static (or) dynamic

● Causality

6
6
LINEARITY PROPERTY:
● Linearity refers to the property of a system where the output
is directly proportional to the input.
● It refers to a fundamental principle in systems and signal
processing where a system is said to be linear if it satisfies the
superposition principle.
Importance of Linearity in 2D Systems:
● Simplification: Linear systems are mathematically simpler
to analyze and solve. Techniques such as superposition make
it easier to predict system behavior.

7 7
● Predictability: In engineering and physics, linear systems
exhibit predictable behavior, making them easier to control
and design.
• Wide Applicability: Many real-world systems can be
approximated as linear systems within a certain range of
operation, enabling the use of linear analysis tools in
practice.
• Efficiency: Linear systems can be efficiently computed using
various algorithms, especially in signal processing and
control systems.

8
8
Superposition Principle:
In a linear system, the response (output) to a combination of
inputs is equal to the sum of the responses to each input
individually.
Conditions for Superposition in 2D Systems:
1. The system should obey Homogeneity and Additivity
property
2. The inputs must be defined in a suitable domain (e.g., time,
space).
3. The system must be time-invariant for consistent results
across different times.

9 9
 Homogeneity:
A system is homogeneous if scaling the input by a
constant factor scales the output by the same factor.
Mathematically, for constant a:
T { a ⋅ x(t1,t2) } = a ⋅T { x(t1,t2) }
Proof for Linearity in Homogeneity in 2D Systems:
Let’s assume x(t1,t2) is an arbitrary input to the system.
● The output for this input is:
y(t1,t2)=T{x(t1,t2)}
● Now consider the scaled input a⋅x(t1,t2):
T{a⋅x(t1,t2)}

10 1
 ● By the property of homogeneity:

T{ a ⋅ x(t1,t2) } = a ⋅ T{ x(t1,t2) } = a ⋅ y(t1,t2)

This demonstrates that scaling the input by a results in the

output being scaled by the same factor, thereby confirming
homogeneity.

11 1
 Additivity:
A system is additive if the response to the sum of two
inputs equals the sum of the responses to each input.
Formally, for inputs x1(t1,t2) and x2(t1,t2):
T{ x1(t1,t2) + x2(t1,t2) } = T{ x1(t1,t2) } + T{ x2(t1,t2) }
Proof for Linearity in Additivity in 2D Systems:
Let’s define:
• y1(t1,t2) = T{ x1(t1,t2) }
• y2(t1,t2) = T{ x2(t1,t2) }

12 1
 ● Now, consider the combined input:
T{x1(t1,t2)+x2(t1,t2)}
● By the property of additivity:
T{ x1(t1,t2) + x2(t1,t2) } = T{ x1(t1,t2) } + T{ x2(t1,t2) }
= y1(t1,t2)+y2(t1,t2)
This proves that the output for the sum of two inputs equals the
sum of the outputs for the individual inputs, confirming the
additivity property.

13 1
 PROBLEM:
1. Consider the following 2D system:
Y(t1,t2)=T{x(t1,t2)}=t1^2.x(t1,t2)+3t2.x(t1,t2)
Solution: Verify the linearity by checking,
1.Homogenetiy
For an input x(t1,t2),the output is,
y(t1,t2)=t1^2.x(t1,t2)+3t2.x(t1,t2)
Test homogeneity by scaling the input by a constant a:
T{a,x(t1,t2)}=t1^2.(a.x(t1,t2))+3t2.(a.x(t1,t2))
Simply the expression:
T{a.x(t1,t2)}=a.(t1^2.x(t1,t2)+3t2.x(t1,t2))=a.T{x(t1,t2)}
T{a.x(t1,t2)} = a.T{x(t1,t2)}
The system satisfies homogeneity.

14 1
 2.Additivity
For two inputs x1(t1,t2) and x2(t1,t2),the output is
y1(t1,t2)=t1^2.x1(t1,t2)+3t2.x1(t1,t2)
y2(t1,t2)=t1^2.x1(t1,t2)+3t2.x1(t1,t2)
Test the additivity by summing the inputs:
T{x1(t1,t2)+x2(t1,t2)}=
t1^2.(x1(t1,t2)+x2(t1,t2))+3t2.(x1(t1,t2)+x2(t1,t2))
Simplify the expression:
T{x1(t1,t2)+x2(t1,t2)}=
t1^2.x1(t1,t2)+3t2.x1(t1,t2)+t1^2.x1(t1,t2)+3t2.x1(t1,t2)
T{x1(t1,t2)+x2(t1,t2)}=T{x1(t1,t2)}+T{x2(t1,t2)}
The system satisfies additivity
Since the system satisfies both homogeneity and additivity, it is linear

15 1
 2. The input and output relationship of the following sytems are
given.
Examine whether the systems are linear or not .
a) y(n1,n2)=x^2(n1,n2)
b) y(n1,n2)=cos(x(n1,n2))

SOLUTION:
For checking the linearity , the system should be in relaxed state.
A system is said to be relaxed if the output is zero when the input is zero.
Formally, if:
● x(n1,n2) = 0 ∀(n1,n2) then, y(n1,n2) = 0 ∀(n1,n2)

16
16
 (a): y(n1,n2)=x^2(n1,n2)
Step 1 : checking the state of the system
y(n1,n2) = x^2(n1, n2)
• If x(n1,n2) = 0, then: y(n1,n2) =(0)^2 = 0
• Therefore, this system is relaxed.
Step 2 : Checking Additivity Property
we need to check if :
T[ x1(n1,n2) + x2(n1,n2) ] = T[ x1(n1,n2) ] + T[ x2(n1,n2) ]
For the left side, the output is:
T[x1(n1,n2) + x2(n1,n2)] = (x1(n1,n2) + x2(n1,n2))^2
Expanding this:
(x1(n1,n2)+x2(n1,n2))^2 =
x1^2(n1,n2)+2x1(n1,n2)x2(n1,n2)+x2^2(n1,n2)
17 17
 • For the right side, the sum of the individual outputs is:
T[x1(n1,n2)]+T[x2(n1,n2)]=x1^2(n1,n2)+x2^2(n1,n2)
• Clearly, the left side contains the additional term
2x1(n1,n2)x2(n1, n2) so:
• (x1(n1,n2)+x2(n1,n2))^2 ≠ x1^2(n1,n2)+x2^2(n1,n2
• Thus, the system fails the additivity test.
Step 3 : Checking Homogeneity property
We need to check if:
T[αx(n1,n2)] = αT[x(n1,n2)]
For the left side, the output is:
T[αx(n1,n2)] = (αx(n1,n2))^2
=α^2 x^2(n1,n2)

18
18
 For the right side, the scaled output is:
αT[x(n1,n2)] = α x^2(n1,n2)
Since α^2 x^2(n1,n2) ≠ α x^2(n1,n2 (except when α=1 or α=0),
the system fails the homogeneity test.
● Conclusion for (a): y(n1,n2)=x^2(n1,n2) is nonlinear.

b) y(n1​,n2​) = cos(x(n1​,n2​))
If x(n1,n2) = 0 then:
y(n1,n2) = cos⁡(0) = 1
Since the output is not zero when the input is zero, this system
is not relaxed.
So we no need to check for linearity property.

19
19
Advantages :
● Extraction

● Noise Reduction

● Image Compression

● Transformation analysis

20
20
 Disadvantages :
● High complexity

● Storage demands

● Processing time

● Noise sensitivity

21
21
 Applications of Linearity:
• Filters in digital signal processing, which are often
linear time-invariant (LTI) systems, follow the linearity
property.
• Fourier Transform: The Fourier transform of a sum of
signals is the sum of the Fourier transforms of each
individual signal, which showcases linearity.

22 2
 CONCLUSION :
● The principles of linearity(additivity and homogeneity) and
the superposition principle extend naturally from 1D to 2D
systems. In both cases, linear systems produce outputs that
respect the sum and scaling of inputs, allowing us to
decompose complex inputs into simpler parts and predict
the overall system behavior based on individual responses.

23 2
 REFERENCES:
1) Two Dimensional Signal Processing , Jae S.Lim
(Massachusetts Institute of Technology)
2) https://course.ece.cmu.edu/~ece491/lectures/L26/
LimChapter.pdf

24
2
 THANK YOU

25 25