WAVELET TRANSFORM

Transform Domain: Transform domain is generally used to represent/analyze the signal characteristics in Different format by changing the independent variable.

A complex signal: The Stationary Signal

x(t)=cos(2*pi*10*t)+cos(2*pi*25*t)+cos(2*pi*50*t)+cos(2*pi*100*t)

Its frequency domain:

Another Complex Signal: Non-Stationary Signal Its Time Domain:

Its Frequency Domain:

The Wavelet Transform: Time-Frequency Representations: f ^ Continuous Wavelet Transform |******************************************* |* * * * * * * * * * * * * * * |* * * * * * * |* * * * |* * --------------------------------------------> time | | Discrete Wavelet Transform | ******************************************************* |* * * * * * * * * * * * * * * * * * |* * * * * * * * * |* * * * * |* * * |----------------------------------------------------------> time

Example of Continuous Wavelet Transform:

Fourier and its inverse transform:

The Short Term Fourier Transform:

Example of Short Term Fourier Transform:

Narrow Window => Good Time Resolution and Poor Frequency Resolution Wide Window => Poor Time Resolution and Good Frequency Resolution Application of windows of different widths:

For a = 0.001

For a = 0.0001

For a = 0.00001

It can be clearly observed that with the increase in the window width, the time resolution is degraded.

The Continuous Wavelet Transform: Developed as an alternative to STFT to avoid resolution problems. Negative frequencies are not computed. The width of the window is changed for each components calculation. The CWT can be defined as below:

Function of two variables Translation (TAU) and Scale (S). Psi(tau,s) is called the mother wavelet. Wavelet means small wave. Mother wavelet is prototype of generating other window functions.

Translation is related to location of the window.

Frequency is replaced by scale.

The Scale: High scales corresponds to non-detailed global view. Smaller scales corresponds to detailed view. High scales => Low Frequencies & Low Scales => High Frequencies. Scaling either dilates or compresses the signal

Computation of CWT: The windows used in the computation are either dilated or compressed versions of mother wavelet. As the real time signals are band limited, only a finite set of intervals of scale are sufficient to calculate. Example for S=1:

Example for S=5:

For every scale and every time interval, a point on the time scale plane is computed. The computations at one scale construct row on time-scale plane and computations at different scales construct columns on time-scale plane. Example Signal:

CWT output:

The Resolution solution:

Time and Frequency Resolutions:

Though the dimensions of the boxes are changing, area remains same.

Mathematical Aspects: Basis Vectors: Every vector in a vector space can be written as a linear combination of the basis vectors in the vector space. A basis of a vector space V is a set of linearly independent vectors, such that any Vector v in V can be written as a linear combination of these basis vectors.

And it can be generalized to functions as,

Inner Products:
Let f(t) and g(t) be two functions in L^2 [a,b]. ( L^2 [a,b] denotes the set of square integrable functions in the interval [a,b]). The inner product of two functions is defined by

According to the above definition of the inner product, the CWT can be thought of as the inner product of the test signal with the basis functions psi_(tau ,s)(t):

where,

Orthogonality: Two vectors v , w are said to be orthogonal if their inner product equals zero:

Similarly, two functions f and g are said to be orthogonal to each other if their inner product is zero:

Orthonormality: A set of vectors {v_1, v_2, ....,v_n} is said to be orthonormal , if they are pairwise orthogonal to each other, and all have length ``1''. This can be expressed as:

Similarly, a set of functions {phi_k(t)}, k=1,2,3,..., is said to be orthonormal if

and or equivalently

where, delta_{kl} is the Kronecker delta function, defined as:

The orthonormal bases allow computation of these coefficients in a very simple and straightforward way using the orthonormality property. For orthonormal bases, the coefficients can be calculated as

and the function f(t) can then be reconstructed by

Some example wavelets:

The Mexican Hat wavelet is defined as the second derivative of the Gaussian function:

which is

The Morlet wavelet is defined as where a is a modulation parameter, and sigma is the scaling parameter that affects the width of the window. The Wavelet Synthesis: CWT is a reversible transform provided the condition for admissibility is satisfied. Inverse Continuous Wavelet Transform:

Admissibility condition:

Discretization of the Continuous Wavelet Transform: The Wavelet Series: Discretized for computational purpose. In the case of WT, the scale change can be used to reduce the sampling rate.

if the time-scale plane needs to be sampled with a sampling rate of N1 at scale s1, the same plane can be sampled with a sampling rate of N2 , at scale s2 , where, s1 < s_2 (corresponding to frequencies f1>f2 ) and N2 < N1 . The actual relationship between N1 and N2 is

Nyquist Criteria for sampling need to be obeyed if reconstruction is important.

The Reconstruction Process from discrete wavelet series:

The scale parameter s is discretized first on a logarithmic grid. The time parameter is then discretized with respect to the scale parameter , i.e., a different sampling rate is used for every scale. In other words, the sampling is done on the dyadic sampling grid shown

The continuous wavelet function

by inserting s = s0^j , and tau = k.s0^j.tau0 . If {psi_(j,k)} constitutes an orthonormal basis, the wavelet series transform becomes

or A wavelet series requires that {psi_(j,k)} are either orthonormal, biorthogonal, or frame. If {psi_(j,k)} are not orthonormal, the above equation becomes,

The Discrete Wavelet Transform: The discrete wavelet transform (DWT) is a representation of a signal x (t) belongs to L2 using an orthonormal basis consisting of a countably-innite set of wavelets. Denoting the wavelet basis as { k,n (t) |k 2 Z ^ n 2 Z}, the DWT transform pair is

where {dk,n} are the wavelet coecients.

In CWT, the signals are analyzed using a set of basis functions which relate to each other by simple scaling and translation. In the case of DWT, a time-scale representation of the digital signal is obtained using digital filtering techniques. The signal to be analyzed is passed through filters with different cutoff frequencies at different scales.

Subband Coding: A time-scale representation of a digital signal is obtained using digital filtering techniques. 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 upsampling and downsampling (subsampling) operations. DWT coefficients are usually sampled from the CWT on a dyadic grid, i.e., s0 = 2 and 0 = 1, yielding s=2j and =k*2j. The procedure starts with passing this signal (sequence) through a half band digital lowpass filter with impulse response h[n]. Filtering a signal corresponds to the mathematical operation of convolution of the signal with the impulse response of the filter. The convolution operation in discrete time is defined as follows:

This procedure can mathematically be expressed as

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.

The above procedure, which is also known as the subband coding, can be repeated for further decomposition. At every level, the filtering and subsampling will result in half the number of samples (and hence half the time resolution) and half the frequency band spanned (and hence double the frequency resolution).

The subband coding is depicted in below figure:

One important property of the discrete wavelet transform is the relationship between the impulse responses of the highpass and lowpass filters. The highpass and lowpass filters are not independent of each other, and they are related by

The reconstruction in this case is very easy since halfband filters form orthonormal bases. Analysis and synthesis filters are identical to each other,except for a time reversal.

Therefore, the reconstruction formula becomes (for each layer)

END