On Two-Dimensional Quaternion Wigner-Ville Distribution

Bahri, Mawardi

doi:https://doi.org/10.1155/2014/139471

Journal of Applied Mathematics

On this page

Abstract Introduction Acknowledgments References Copyright Related Articles

Research Article | Open Access

Volume 2014 | Article ID 139471 | https://doi.org/10.1155/2014/139471

On Two-Dimensional Quaternion Wigner-Ville Distribution

Mawardi Bahri¹

Academic Editor: Sabri Arik

Received04 Aug 2013

Revised24 Dec 2013

Accepted24 Dec 2013

Published22 Jan 2014

Abstract

We present the two-dimensional quaternion Wigner-Ville distribution (QWVD). The transform is constructed by substituting the Fourier transform kernel with the quaternion Fourier transform (QFT) kernel in the classical Wigner-Ville distribution definition. Based on the properties of quaternions and the QFT kernel we obtain three types of the QWVD. We discuss some useful properties of various definitions for the QWVD, which are extensions of the classical Wigner-Ville distribution properties.

1. Introduction

The classical Wigner-Ville distribution (WVD) or Wigner-Ville transform (WVT) is an important tool in the time-frequency signal analysis. It was first introduced by Eugene Wigner in his calculation of the quantum corrections of classical statistical mechanics. It was independently derived again by J. Ville in 1948 as a quadratic representation of the local time-frequency energy of a signal. In [1–3], the authors introduced the WVT and established some important properties of the WVT. The transform is then extended to the linear canonical transform (LCT) domain by replacing the kernel of the classical Fourier transform (FT) with the kernel of the LCT in the WVD domain [4].

As a generalization of the real and complex Fourier transform (FT), the quaternion Fourier transform (QFT) has been of interest to researchers for some years. A number of useful properties of the QFT have been found including shift, modulation, convolution, correlation, differentiation, energy conservation, uncertainty principle, and so on. Due to the noncommutative property of quaternion multiplication, there are three different types of two-dimensional QFTs. These three QFTs are so-called a left-sided QFT, a right-sided QFT, and a two-sided QFT, respectively (see, e.g., [5–9]). In [10, 11], special properties of the asymptotic behaviour of the right-sided QFT are discussed and generalization of the classical Bohner-Millos theorems to the framework of quaternion analysis is established. Many generalized transforms are closely related to the QFTs, for example, the quaternion wavelet transform, fractional quaternion Fourier transform, quaternion linear canonical transform, and quaternionic windowed Fourier transform [12–18]. Based on the QFTs, one also may extend the WVD to the quaternion algebra while enjoying similar properties as in the classical case.

Therefore, the main purpose of this paper is to propose a generalization of the classical WVD to quaternion algebra, which we call the quaternion Wigner-Ville distribution (QWVD). Our generalization is constructed by substituting the kernel of the FT with the kernel of the QFT in the classical WVT definition. Due to the non-commutative rule of quaternions and the QFT kernel we obtain the definition of different types of the QWVD. We then derive some important properties of the QWVD such as shift, reconstruction formula, modulation, and orthogonality relation in detail. We present an example to show the difference between the QWVD and the WVD.

The organization of the paper is as follows. The remainder of this section introduces some notations and briefly recalls some general definitions and basic properties of quaternion algebra and quaternion Fourier transform. In Section 3, we provide the basic ideas for the construction of the two-sided QWVD and derive several its important properties using the two-sided QFT. The construction of the right-sided QWVD is provided in Section 4. In Section 5 we introduce quaternion ambiguity function (QAF); its important properties are also discussed in this section.

2. Basics

2.1. Quaternion Algebra

The quaternion algebra was formally introduced by Hamilton in 1843, and it is a generalization of complex numbers. The quaternion algebra over , denoted by , is an associative non-commutative four-dimensional algebra: which obey Hamilton’s multiplication rules

The quaternion conjugate of a quaternion is given by and it is an anti-involution; that is, From (3) we obtain the norm of defined as It is not difficult to see that Using the conjugate (3) and the modulus of , we can define the inverse of as which shows that is a normed division algebra.

We use parenthesis to denote the inner product of two quaternion functions, , as follows: When we obtain the associated norm As a consequence of the inner product (8) we obtain the quaternion Cauchy-Schwarz inequality

2.2. Quaternion Fourier Transform (QFT)

The quaternion (or hypercomplex) Fourier transform is defined similar to the classical FT of the 2D functions. The noncommutative property of quaternion multiplication allows us to have three different definitions of the QFT. In the following we briefly introduce the two-sided QFT and the right-sided QFT. For more details we refer the reader to [7, 8, 19].

Definition 1 (two-sided QFT). The two-sided QFT of is the transform given by the integral where , , and the quaternion exponential product is called the quaternion Fourier kernel.

Theorem 2 (inverse two-sided QFT). Suppose that and . Then the two-sided QFT of is an invertible transform and its inverse is given by where the quaternion exponential product is called the inverse two-sided quaternion Fourier kernel.

Notice that if from the inverse of the two-sided QFT we obtain quaternion Dirac’s delta function; that is,

We have known that Parseval’s formula is not valid for the two-sided QFT, but a special case of Parseval’s formula called the Plancherel formula remains valid; that is, We introduce Parseval’s formula for the right sided QFT as where the right-sided QFT is defined by for every

2.3. Fundamental Operators

Before we discuss the QWVD we need to introduce some notations, which will be used in the next section. For a quaternion function we define the translation, modulation, and dilation as follows: where and . The composition of the translation and modulation is called the time-frequency shift; that is, Just as the classical case (see [3]), we obtain the canonical commutation relations

3. Quaternion Wigner-Ville Distribution (QWVD)

In this section, we introduce the 2D quaternion Wigner-Ville distribution (QWVD). We investigate several basic properties of the QWVD which are important for signal representation in signal processing.

3.1. Definition of 2-D QWVD

Based on the properties of quaternions and the definition of the classical Wigner-Ville transform associated with the Fourier transform, we obtain a definition of the QWVD by replacing the kernel of the FT with the kernel of the two-sided QFT in the classical WVD definition as follows.

Definition 3. The cross two-sided quaternion Wigner-Ville distribution of two-dimensional functions (or signals) is given by provided the integral exists.

It should be remembered that the kernel of the cross two-sided QWVD in (20) does not commute with quaternion functions and so that several properties of the WVD are not valid in the cross two-sided QWVD.

By making the change of variables , (20) can be written in the form We see that the above expression gives an equivalent definition of in the form where . Here is the windowed quaternionic Fourier transform which was recently proposed by Fu et al. [13]. Furthermore, if we write , we immediately obtain which tells us that the cross two-sided QWVD is in fact the two-sided QFT of the function with respect to . This fact is very important in proving Moyal’s formula for the two-sided QWVD.

Lemma 4. Let be two quaternion-valued functions. Then is bounded on ; that is, In particular, if , the above expression reduces to

Proof. By application of the quaternion Cauchy-Schwarz inequality (10) we easily obtain It means that for all and we have which was to be proved.

The following theorem shows that the cross two-sided QWVD is invertible; that is, the original quaternion signal can be uniquely determined in terms of its cross two-sided QWVD within a constant factor.

Theorem 5 (reconstruction formula for the two-sided QWVD). The inverse transform of the cross two-sided QWVD of the signal is given by provided .

Proof. From the definition of the cross two-sided QWVD, we know that Indeed, from the inverse of the two-sided QFT (12), it follows that Letting , the above expression will lead to and the final result can be obtained by letting ; that is, which completes the proof.

Lemma 6. For any and , if and , then we obtain

Proof. Since is the two-sided QFT of , (14) yields Using the change of variables and and then integrating (34) with respect to we immediately get This finishes the proof.

Remark 7. Equation (33) is known as the radar uncertainty principle in the cross two-sided QWVD domain. It shows that the quaternion function cannot be concentrated arbitrarily close to the origin.

We obtain the following results which correspond to classical WVD properties (compare to [1, 2]).

Lemma 8. For , the two-sided QFT of (20) with respect to can be represented in the form

Proof. A simple computation gives for every where as usual denotes the Dirac delta function (13).

It is not difficult to see that for we get the auto two-sided quaternion Wigner-Ville distribution defined by Both the cross two-sided quaternion Wigner-Ville distribution and the auto two-sided Wigner-Ville distribution are often so-called the two-sided Wigner-Ville distribution or two-sided quaternion Wigner-Ville transform (QWVT). If is a real function, then the change of variables to (38) yields

Lemma 9 (time energy density). If , then the two-sided QWVD satisfies the time energy density as

Proof. A direct calculation shows that Changing variables and in the above expression, we immediately get In particular, if we substitute in (42), we obtain If , then integrating (43) with respect to gives the final result This gives the desired result.

Remark 10. Unlike classical case, we cannot establish the frequency energy density, because the Parseval formula does not hold for the two-sided QFT.

For an illustrative purpose, we consider an example of the two-sided QWVD.

Example 11. Given a Gaussian signal defined by find the two-sided QWVD of a pure sine wave .

From Definition 3 we obtain

3.2. Basic Properties of Two-Sided QWVD

The following propositions describe the elementary properties of the two-sided QWVD and their proofs in detail. We find that most of the expected properties of the classical WVD are still valid with some modifications in this case.

Proposition 12 (shift). Let be quaternion signals. If we time shift the signals by , then we obtain

Proposition 13 (nonlinearity). Let be quaternion signals. The two-sided QWVD is nonlinear. That is

Proof. Applying (20) and the basic properties of quaternions we get The proof is complete.

This shows that the two-sided QWVD of the sum of two quaternion signals is not simply the sum of the two-sided QWVD of the signals.

Proposition 14 (modulation). Let be quaternion signals. If we spectrum-shift the signal by , then we obtain

Proof. In this proof we will use the decomposition of quaternion functions. We first remember the fact that and so on. This further leads to which completes the proof.

Proposition 15 (dilation). Let be quaternion signals. Then

Proof. Let , then a direct computation yields which was to be proved.

3.3. Main Properties of Two-Sided QWVD

In this subsection we investigate two main properties of the two-sided QWVD. Based on the properties of the two-sided QFT we can only establish the specific Moyal formula for the QWVD as follows.

Theorem 16 (Moyal’s formula for the two-sided QWVD). Let be two quaternion signals. Then the following equation holds:

Proof. We first notice that, for fixed , Applying (55) and Parseval’s formula (14) we get Integrating (56) with respect to we further obtain First use , then change the order of integration and use we immediately get This is the desired result.

Due to the non-commutativity of the quaternion exponential products and quaternion multiplications, we only can establish a special condition of the convolution theorem for the two-sided QWVD (compared to [20]). This is described in the following theorem.

Theorem 17 (convolution for the two-sided QWVD). Let be two real-valued signals. If we assume that is a real-valued function, then the following result holds: where is the quaternion convolution operator.

Proof. Applying the definition of the two-sided QWVD (20) and elementary properties of the quaternion convolution gives By the change of variables and , the above identity can be written as Because are real-valued signals, we may interchange the order of and to get where in the third line we used the assumption to interchange the order of the two-sided QWVD and the kernel of the two-sided QFT. This finishes the proof of the theorem.

4. Right-Sided QWVD

Based on the properties of quaternions and the kernel of the right-sided QFT we may construct the right-sided QWVD. We will see that some properties of the right-sided QWVD are quite different from the two-sided QWVD.

Definition 18. The cross right-sided quaternion Wigner-Ville distribution of the 2D signals is given by provided the integral exists.

Similar to inverse transform of the two-sided QWVD we get the following fundamental result.

Theorem 19. The inverse right-sided QWVD of the signal is given by provided .

By using the change of variables , (63) can be expressed in the following form: Observe first that if and are real signals, the above expression gives an equivalent definition of in the form where . Here is the windowed quaternionic Fourier transform which was introduced by Bahri et al. [17, 18].

Lemma 20 (time energy density). If , then the right-sided QWVD satisfies the time energy density as

Proof. Indeed, we have Integrating both sides of the above expression with respect to , we obtain In particular, , we get This gives the desired result.

4.1. Useful Properties

Proceeding as in the proof of propositions listed in Section 3.2 we obtain elementary properties of the right-sided QWVD.

Proposition 21. For the right-sided QWVD has the following useful properties.
(i) Shift. One has
(ii) Nonlinearity. One has
(iii) Dilation. One has

Remark 22. It seems that modulation property is not valid for the right-sided QWVD. This shows that some properties of the right-sided QWVD follow the properties the right-sided QFT.

In the following we establish general Moyal’s formula of the right-sided QWVD. We see that Moyal’s formula of the two-sided QWVD is a special case of Moyal’s formula of the right-sided QWVD.

Theorem 23 (Moyal’s formula for right-sided the QWVD). Let be quaternion-valued signals. Then the following equation holds:

Proof. From the definition of the right-sided QWVD (63), we easily obtain Letting , and putting , Interchanging the order of integration gives This completes the proof of the theorem.

Based on the above theorem, we may conclude the following important consequences.(i) If , then

This formula is quite similar to Moyal’s formula for the classical WVD, for example; see [1, 3]. However, we must remember that (78) is a quaternion-valued function.(ii)If , then (iii)If and , then

This formula has the same form as the specific Moyal formula for the two-sided QWVD (54) and also the classical WVD.

5. Quaternion Ambiguity Function (QAF)

The classical ambiguity function (AF) is firstly introduced by Woodward in 1953 for mathematical analysis of sonar and radar signals [1]. This section will generalize the classical AF in the quaternion algebra setting.

Definition 24. The cross two-sided quaternionic ambiguity function (QAF) of the two-dimensional functions (or signals) is denoted by and is defined by provided the integral exists.

The following lemma describes the relationship between the two-sided QWVD and the two-sided QAF mentioned above.

Lemma 25. The two-sided QWVD of the signal can be seen as the two-sided QAF of the signals by formula

Proof. Putting , we may write (81) in the form It means that we have We should remember that if we write , then the two-sided QWVD defined in (81) is the two-sided QFT of the function with respect to . That is, which was to be proved.

We also obtain the following results which correspond to classical WVD properties (compared to [1, 2]).

Lemma 26. For , the two-sided QFT of (81) with respect to can be represented in the form

Proof. Direct calculations yield which was to be proved.

The following theorem shows that the quaternion signal can be recovered from the two-sided QAF up to a quaternion constant.

Theorem 27 (reconstruction formula for two-sided QAF). The inverse transform of the cross two-sided QWVD of the signal is given by provided .

Proof. We have from the inverse transform of the two-sided QFT (12) Taking the specific value, , we have Or, equivalently, which completes the proof.

5.1. Useful Properties of Two-Sided QAF

The properties of the two-sided QAF are summarized in the following proposition. It seem that they have a remarkable similarity with those of the two-sided QWVD.

Proposition 28. For the two-sided QAF has the following important properties. (i)Shift. One has (ii)Nonlinearity. One has (iii)Modulation. One has (iv)Dilation. One has

Proof. For part (i), using the definition of the two-sided QAF (81), we obtain To derive part (iii), we apply (81) and the decomposition of quaternion function to get For part (iv), let , we have, by definition,

Theorem 29 (specific Moyal’s formula for the two-sided QAF). Suppose that and are two quaternion signals. Then the following equation holds:

Proof. Using Plancherel’s formula for the right-sided QFT in (14) and then applying Lemma 26 we immediately obtain Integrating (100) with respect to yields Putting and , then the right side of the above expression can be rewritten as After switching the order of integration, we get This is the desired result.

Conflict of Interests

The author declares that there is no conflict of interests regarding the publication of this paper.

Acknowledgments

The author would like to thank the anonymous reviewer for valuable comments and suggestions which improved the clarity and quality of this paper. This work is partially supported by Hibah Penelitian Kompetisi Internal 2013 (no. 110/UN4-.42/LK.26/SP-UH/2013) from the Hasanuddin University, Indonesia.

References

L. Debnath, Wavelet Transforms and Their Applications, Birkhäuser, Boston, Mass, USA, 2002.
View at: Publisher Site | MathSciNet
L. Debnath, B. V. Shankara, and N. Rao, “On new two-dimensional Wigner-Ville nonlinear integral transforms and their basic properties,” Integral Transforms and Special Functions, vol. 21, no. 3, pp. 165–174, 2010.
View at: Publisher Site | Google Scholar | Zentralblatt MATH | MathSciNet
K. Gröchenig, Foundations of Time-Frequency Analysis, Birkhäuser, Boston, Mass, USA, 2001.
View at: MathSciNet
R.-F. Bai, B.-Z. Li, and Q.-Y. Cheng, “Wigner-Ville distribution associated with the linear canonical transform,” Journal of Applied Mathematics, vol. 2012, Article ID 740161, 14 pages, 2012.
View at: Publisher Site | Google Scholar | Zentralblatt MATH | MathSciNet
T. Bülow, Hypercomplex spectral signal representations for the processing and analysis of images [Ph.D. thesis], University of Kiel, Kiel, Germany, 1999.
T. Bülow, M. Felsberg, and G. Sommer, “Non-commutative hypercomplex Fourier transforms of multidimensional signals,” in Geometric Computing with Clifford Algebras—Theoretical Foundations and Applications in Computer Vision and Robotics, G. Sommer, Ed., pp. 187–207, Springer, Berlin, Germany, 2001.
View at: Publisher Site | Google Scholar | Zentralblatt MATH | MathSciNet
E. M. S. Hitzer, “Quaternion Fourier transform on quaternion fields and generalizations,” Advances in Applied Clifford Algebras, vol. 17, no. 3, pp. 497–517, 2007.
View at: Publisher Site | Google Scholar | Zentralblatt MATH | MathSciNet
M. Bahri, E. S. M. Hitzer, A. Hayashi, and R. Ashino, “An uncertainty principle for quaternion Fourier transform,” Computers & Mathematics with Applications, vol. 56, no. 9, pp. 2398–2410, 2008.
View at: Publisher Site | Google Scholar | Zentralblatt MATH | MathSciNet
M. Bahri, R. Ashino, and R. Vaillancourt, “Convolution theorems for quaternion Fourier transform: properties and applications,” Abstract and Applied Analysis, vol. 2013, Article ID 162769, 10 pages, 2013.
View at: Publisher Site | Google Scholar
S. Georgiev and J. Morais, “Bochner's theorems in the framework of quaternion analysis,” in Quaternion and Clifford Fourier Transforms and Wavelets, E. Hitzer and S. Sangwine, Eds., Trends in Mathematics, pp. 85–104, Springer, Basel, Switzerland, 2013.
View at: Publisher Site | Google Scholar
S. Georgiev, J. Morais, K. I. Kou, and W. Sprössig, “Bochner-Minlos theorems and quaternion Fourier transform,” in Quaternion and Clifford Fourier Transforms and Wavelets, E. Hitzer and S. Sangwine, Eds., Trends in Mathematics, pp. 105–120, Springer, Basel, Switzerland, 2013.
View at: Publisher Site | Google Scholar
H. De Bie and N. De Schepper, “Fractional Fourier transforms of hypercomplex signals,” Signal, Image and Video Processing, vol. 6, no. 3, pp. 381–388, 2012.
View at: Publisher Site | Google Scholar
Y. Fu, U. Kähler, and P. Cerejeiras, “The Balian-low theorem for the windowed quaternionic Fourier transform,” Advances in Applied Clifford Algebras, vol. 22, no. 4, pp. 1025–1040, 2012.
View at: Publisher Site | Google Scholar | Zentralblatt MATH | MathSciNet
K.-I. Kou, R.-H. Xu, and Y.-H. Zhang, “Paley-Wiener theorems and uncertainty principles for the windowed linear canonical transform,” Mathematical Methods in the Applied Sciences, vol. 35, no. 17, pp. 2122–2132, 2012.
View at: Publisher Site | Google Scholar | Zentralblatt MATH | MathSciNet
K. I. Kou, J.-Y. Ou, and J. Morais, “On uncertainty principle for quaternionic linear canonical transform,” Abstract and Applied Analysis, vol. 2013, Article ID 725952, 14 pages, 2013.
View at: Publisher Site | Google Scholar | Zentralblatt MATH | MathSciNet
M. Bahri, “A generalized windowed Fourier transform in real Clifford algebra ${C l}_{0, n}$ ,” in Quaternion and Clifford Fourier Transforms and Wavelets, E. Hitzer and S. Sangwine, Eds., Trends in Mathematics, pp. 285–298, Springer, Basel, Switzerland, 2013.
View at: Publisher Site | Google Scholar
M. Bahri, E. S. M. Hitzer, R. Ashino, and R. Vaillancourt, “Windowed Fourier transform of two-dimensional quaternionic signals,” Applied Mathematics and Computation, vol. 216, no. 8, pp. 2366–2379, 2010.
View at: Publisher Site | Google Scholar | Zentralblatt MATH | MathSciNet
M. Bahri, “A generalized windowed Fourier transform for quaternions,” Far East Journal of Applied Mathematics, vol. 42, no. 1, pp. 35–47, 2010.
View at: Google Scholar | Zentralblatt MATH | MathSciNet
S. J. Sangwine and T. A. Ell, “Hypercomplex Fourier transforms of color images,” IEEE Transactions on Image Processing, vol. 16, no. 1, pp. 22–35, 2007.
View at: Publisher Site | Google Scholar | MathSciNet
M. Bahri, “Correlation theorem for Wigner-Ville distribution,” Far East Journal of Mathematical Sciences, vol. 80, no. 1, pp. 123–133, 2013.
View at: Google Scholar

Copyright

Copyright © 2014 Mawardi Bahri. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

PDF Download Citation

Download other formats

Order printed copies

Views

1719

Downloads

1325

Citations