Wavelet Transform

Presented By
Prof. Arshi Khan
 Why transform?
• Mathematical transformations are applied to signals to obtain
a further information from that signal that is not readily
available in the raw signal.

• The Fourier transforms are probably by far the most popular.

• Most of the signals in practice, are TIME-DOMAIN signals in

their raw format. That is, whatever that signal is measuring,
is a function of time.
 Why transform?
• When we plot time domain signals, we obtain a time-
amplitude representation of the signal.

• In many cases, the most distinguished information is hidden in

the frequency content of the signal.

• Frequency spectrum is basically the no of frequencies present

in a given signal.
The Fourier transform
 Why not Fourier Transform ?
• FT of a signal in time domain is taken, the frequency-
amplitude representation of that signal is obtained.

• FT (as well as WT) is a reversible transform, that is, it allows to

go back and forward between the raw and processed
(transformed) signals. It tells us how much of each frequency
exists in the signal, but it does not tell us when in time these
frequency components exist.

• This information is not required when the signal is so-called

stationary .
FT of a Stationary signal
FT of a non-stationary signal
 Short Time Fourier transform
• Transforms such as wavelet transform, STFT, Wigner
distribution etc provides the time-frequency representation of
the signal.
• Wavelet transform was developed to overcome resolution
related problems of STFT. Signal is divided into small enough
segments, where these segments (portions) of the signal can
be assumed to be stationary. For this purpose, a window
function "w" is chosen. The width of this window must be
equal to the segment of the signal where its stationarity is
valid.
 Short Time Fourier Transform: Steps
• The window function is first located to the very beginning of
the signal.
• Let the width of the window be "T" s.
• At this time instant (t=0), the window function will overlap
with the first T/2 seconds.
• The window function and the signal are then multiplied. By
doing this, only the first T/2 seconds of the signal is being
chosen, with the appropriate weighting of the window. Then
this product is assumed to be just another signal, whose FT is
to be taken.
 STFT: Steps(contd….)
• The result of this transformation is the FT of the first T/2
seconds of the signal.

• The next step, would be shifting this window (for some t1

seconds) to a new location, multiplying with the signal, and
taking the FT of the product.

• This procedure is followed, until the end of the signal is

reached by shifting the window with "t1" seconds interval.
 STFT(contd…….)
• The following definition of the STFT summarizes all the above
explanations in one line:

• For every t' and f a new STFT coefficient is computed.

STFT of a Non-Stationary signal
 Problems with STFT
• Heisenberg’s Uncertainty principle: one cannot know the
exact time-frequency representation of a signal. one can
know the time intervals in which certain band of
frequencies exist, which is a resolution problem.
• The problem with the STFT is width of the window function
that is used.
• window is of finite length, it covers only a portion of the
signal, which leads to poor frequency resolution . Exact
frequency components that exist in the signal are not known,
but we only know a band of frequencies that exist.
 Problems with STFT(contd…)
• Narrow window ===>good time resolution, poor frequency
resolution.

• Wide window ===>good frequency resolution, poor time

resolution.

 The above two points are illustrated with the help of following
example.

• STFT of the signal shown below is calculated using a Gaussian

function of the form : w(t)=exp(-a*(t^2)/2)
Problems with STFT(contd…)
Problems with STFT(contd…)
 Problems with STFT(contd…)
• STFT for a=0.01
 Problems with STFT(contd…)
• STFT for a= 0.0001
 Problems with STFT(contd…)
• STFT for a= 0.00001
Multiresolution Analysis & Continuous
Wavelet Transform
• MRA analyzes the signal at different frequencies with
different resolutions.

• MRA is designed to give good time resolution and poor

frequency resolution at high frequencies and good frequency
resolution and poor time resolution at low frequencies.
Multiresolution Analysis & Continuous
Wavelet Transform(Contd…)
 What is Wavelet analysis?
• Wavelet: It is a small wave with finite energy.
• It allows simultaneous time & frequency analysis.
 Continuous Wavelet transform
• Developed as an alternative to STFT.

• Used to overcome resolution problem of STFT.

• Defined as the sum over all time of the signal multiplied by

the scaled and shifted versions of the wavelet function 𝛹.

• Result of CWT are wavelet coefficients which are function of

scale and position( i.e. translation)

• In CWT the width of the window is changed as the

transform is computed for every single spectral component.
Continuous Wavelet transform(contd..)
• CWT is defined mathematically as:
Continuous Wavelet transform(contd..)

• Mother Wavelet

– A prototype for generating the other window functions

– All the used windows are its dilated or compressed and

shifted versions.
 SCALE
• Scale
– S>1: dilates the signal
– S<1: compresses the signal
• Low Frequency -> High Scale -> Non-detailed Global View of
Signal -> Span Entire Signal
• High Frequency -> Low Scale -> Detailed View of a hidden
pattern in a signal that lasts for a short time.
• Only Limited Interval of Scales is Necessary
 SCALE (contd…)
• In the following figure, s=0.05 is the smallest scale, and s=1 is
the largest scale.
 Computation of CWT
• Step 1: The wavelet is placed at the beginning of the signal,
and set s=1 (the most compressed wavelet)

• Step 2: The wavelet function at scale “1” is multiplied by the

signal, and integrated over all times; then multiplied by a
constant

• Step 3: wavelet is shifted to the right by tau with s=1 and

transformed value is calculated.
 Computation of CWT
• Step 4: This procedure is repeated until the wavelet reaches
the end of the signal

• Step 5: Scale ‘s’ is increased by a sufficiently small value, the

above procedure is repeated for every value of ‘s’.

• Step 6: Each computation for a given s fills the single row of

the time-scale plane;

• Step 7: CWT is obtained once process is completed for all ‘s’.

 Computation of CWT(contd…)
• Four distinct location of wavelet function are shown in fig
below for ‘s’=1.
Interpretation of Time & Frequency
Resolution
 Example :Comparison of CWT of a
normal person and Alzheimer patient
• Following signal belongs to a normal person.
 Example contd…
• Figure below plots an event related potential of a patient
diagnosed with Alzheimer's disease.
 Discrete Wavelet Transform
• The foundations go back to 1976 when Croiser, Esteban, and
Galand devised a technique to decompose discrete time
signals.

• A time-scale representation of a digital signal is obtained

• using digital filtering techniques.

• In the discrete case, filters of different cutoff frequencies are

used to analyze the signal at different scales.

.
 Discrete Wavelet Transform(contd..)
• The resolution of the signal, which is a measure of the
amount of detail information in the signal, is changed by the
filtering operations, and the scale is changed by up sampling
and down sampling (sub sampling) operations.
• Sub sampling a signal corresponds to reducing the sampling
rate, or removing some of the samples of the signal.
• Up sampling a signal corresponds to increasing the sampling
rate of a signal by adding new samples to the signal.
 DWT Computation: Subband Coding

• The DWT analyzes the signal at different frequency bands

with different resolutions by decomposing the signal into a
coarse approximation and detail information.

• DWT employs two sets of functions, called scaling functions

and wavelet functions which are associated with low pass and
highpass filters, respectively.

• The decomposition of the signal into different frequency

bands is simply obtained by successive highpass and lowpass
filtering of the time domain signal.
 DWT Computation: Subband Coding

• The original signal x[n] is first passed through a half band high
pass filter g[n] and a low pass filter h[n].

• After the filtering, half of the samples can be eliminated

according to the Nyquist’s rule, This constitutes one level of
decomposition.
• mathematically be expressed as follows:
Subband Coding(Contd….)
Data Reduction using DWT
 References
• Robi Polikar, The Wavelet Tutorial,
http://users.rowan.edu/~polikar/WAVELETS/WTpart1.html
• Robi Polikar, Multiresolution Wavelet Analysis of Event
Related Potentials for the Detection of Alzheimer's Disease,
Iowa State University, 06/06/1995
• Gonzalez and Woods, Digital Image Processing, 3rd Edition
• Dhananjay theckedath, Digital Image Processing,MU 2012
Edition.