1. Introduction
The general two-sided quaternion Fourier transform was firstly proposed by Hitzer in [
1,
2]. As we all know, it is a generalized form of the two-sided quaternion Fourier transform. Hitzer further has defined the two-sided quaternion Fourier transform based on the orthogonal planes split and demonstrated some fundamental properties of this transform. In [
3], Hitzer also extended general two-sided quaternion Fourier transform within the context of Clifford algebra, named general two-sided Clifford Fourier transform, and derived its convolution and correlation theorems.
In recent years, many researchers have shown interest in a kind of new signal representation tool, known as the quaternion linear canonical transformations (QLCT). It is well known that the QLCT is a nontrivial generalization of the classical linear canonical transform [
4,
5]. It also can be regarded as the generalization of the quaternion Fourier transform (QFT) because the QFT is a special case of the QLCT. Several useful properties of the QLCT have been extensively studied (see, e.g., [
6,
7,
8,
9,
10,
11] and references therein). As in the quaternion Fourier transform case [
12,
13,
14,
15,
16,
17,
18], there are three different kinds of two-dimensional quaternion linear canonical transforms (QLCTs). They are called left-sided QLCT, right-sided QLCT, and two-sided QLCT, respectively.
In this paper, based on the symmetric decomposition of the two-sided quaternion Fourier transform (QFT), we first provide the derivation of a variation on the Heisenberg-type uncertainty principle related to this transformation. The uncertainty principle describes the interaction between a quaternion function and its QFT. To achieve this, we recall the component-wise uncertainty principle for the QFT and show that it is a special case of the proposed uncertainty principle. We then propose general two-sided quaternion linear canonical transform. We further provide a definition of its convolution operator. This definition is constructed by combining convolution definitions for the classical linear canonical transform and the quaternion Fourier transform. We establish its convolution theorem, which give a significant result of general two-sided quaternion linear canonical transform. This theorem is inspired by the work of Hitzer [
19] who established the convolution theorem associated with general two-sided quaternion Fourier transform. Paper [
20,
21] proposed similar work for convolution theorem, but different type of the quaternion linear canonical transform which established the convolution theorem for the two-sided quaternion linear canonical transform. We finally derive correlation theorem of continuous quaternion signals associated with the general two-sided quaternion linear canonical transform.
The main content of this article is as follows. In
Section 2, we recall the basic knowledge of quaternion algebra and orthogonal planes split that will be needed during the paper.
Section 3 is devoted to the derivation of a variation on the Heisenberg-type uncertainty principle related to the two-sided quaternion Fourier transform. In
Section 4 we recall the general two-sided quaternion Fourier transform. In
Section 5 we discuss convolution definition for general two-sided quaternion linear canonical transform and obtain its convolution theorem. In
Section 6, a correlation theorem related to general two-sided quaternion linear canonical transform is presented. In
Section 7 an application of general two-sided quaternion linear canonical transform is studied. Lastly, the summary of this article is included in
Section 8.
2. Quaternions
For the basic notations and definitions on quaternion algebra, see [
22,
23,
24]. Quaternions are hypercomplex numbers, which requires an associative noncommutative four-dimensional algebra. They can be expressed as
where the elements
, and
have properties:
For every quaternion , the scalar and vector parts of h are denoted by and , respectively.
The conjugate element
is given by
which fulfills
From (
2) we obtain the norm of a quaternion
h in the form
The modulus of a product of two quaternion obeys the property
It is not difficult to see that
Applying the conjugate (
2) and the norm of
h gives the inverse of nonzero quaternion
h as
The inner product for two quaternion-valued functions
is defined as
for all
. In particular, for
, we obtain
For
(
6) becomes
Following Hitzer’s work [
1], we introduce the definition of orthogonal 2D planes split (OPS) with respect to any two pure unit quaternions
as follows
Definition 1. Let be an arbitrary pair of pure quaternions , including the cases . For any we introduce the OPS split parts with respect to the two pure unit quaternion as Moreover one has for and Observe that for
and
Equation (
9) will lead to
3. A Variation on Heisenberg’s Inequality for Quaternion Fourier Transform
We start this part by introducing the definition of the two-sided quaternion Fourier transform (QFT) and present its useful properties. We use these results to obtain one of the main results in this study.
Definition 2. Given h in . We defineWe call the two-sided quaternion Fourier transform of h. Definition 3. For any for which , its inverse is defined by Now we put
and
on the right-hand sides of (
12) and (
13), one can easily obtain
and
By (
14) the decomposition of the quaternion function
f will lead to
In the symmetric form, the above identity may be rewritten as
Now we define the module of
as
Furthermore, we obtain the
-norm
Lemma 1. Suppose that . If with exists and belongs to , then for every we haveand Proof. For
applying (
15) results in
and
By repeating this process
additional times, we obtain (
20). Using a similar argument as Equation (
20), we can obtain the proof for Equation (
21). □
By Riesz’s interpolation theorem, we obtain the Hausdorff–Young inequality (see [
25]), that is
where
with
. Taking the inversion formula of the QFT on both sides of (
24), one has
Below we state and prove the uncertainty principle associated with the two-sided quaternion Fourier transform. Firstly, we recall the component-wise uncertainty principle for the QFT that will be established in next theorem.
Theorem 1. Suppose that and that exists. Then Remark 1. It should be noticed that Theorem 1 is valid for all types of the QFT.
It can be observed that for
we may change
- norm to
- norm on left-hand side of (
26) and obtain the next result.
Theorem 2. Under the conditions as above, we have Proof. Due to (
4) and the Holder’s inequality, we have
By virtue of the Hausdorff–Young inequality (
25) we obtain
For
, we see from (
20) that
For
, we can apply similar arguments as above using (
21), and obtain
This ends the proof of the theorem. □
Remark 2. The non-commutativity of the QFT kernel implies that Theorem 2 is slightly different to the right-sided QFT (compare to Theorem 12 of [26]). Below, Theorem 3 is not valid for the right-sided QFT and left-sided QFT. Theorem 3. If and exists and is also in , thenfor . Proof. Simple computation shows that
The proof is complete. □
It can be observed that for
Theorem 3 changes to
which is directional uncertainty principle for the two-sided quaternion Fourier transform and right-sided quaternion Fourier transform [
26].
4. General Two-Sided Quaternion Fourier Transform
In the following we introduce general two-sided quaternion Fourier transform, which is taken from some papers by Hitzer [
2,
3].
Definition 4. For any , the general two-sided quaternion Fourier transform of a quaternion function h with respect to two pure quaternions such that is defined asprovided that the integral exists. Especially, when
and
then (
30) becomes
which is two-sided quaternion Fourier transform defined in (
12).
For any
the general quaternion Fourier transform mentioned above can be inverted by
From linearity of the integral (
30) we obtain the general two-sided quaternion Fourier transform for the OPS split
as
Definition 5. Suppose that and . The general quaternion Fourier transform for the with respect to two linearly independent unit quaternions and is defined by By using relation (
9), the above identity can be rewritten in the form
5. General Two-Sided Quaternion Linear Canonical Transform
Because general two-sided QLCT is a generalization of two-sided QLCT, many useful properties of two-sided QLCT can be extended for general two-sided QLCT such as linearity, time shift, frequency shift, energy conservation, and uncertainty principles with some modifications. Another very important property of general two-sided QLCT is the convolution theorem. We first provide a definition of general two-sided QLCT and its relation to two-sided QLCT. We also present a theorem which describes the relationship between the orthogonal 2D plane split and general two-sided QLCT.
5.1. Definition of General Two-Sided QLCT
Definition 6 (General Two-sided QLCT).
Suppose that and . For any , the general two-sided quaternion linear canonical transform of a quaternion function h with respect to two pure quaternions such that is defined aswhere the kernel functions of the general two-sided QLCT above are given byand On the condition that the general two-sided QLCT parameters satisfy
or
, the general two-sided QLCT of a signal is essentially a quaternion chirp multiplication and is of no particular interest for our objective interests. Therefore, without loss of generality we focus mainly on the general two-sided QLCT in the case of
. For specific parameter matrices
with
, the general two-sided QLCT definition (
37) is reduced to general two-sided QFT definition, that is,
The inversion formula of the general two-sided QLCT is given by
provided that the integral exists.
Based on Definition 5 we can easily obtain the following important theorem, which will be required to derive the main result in the sequel.
Theorem 4 (compare to [
27]).
Let be a quaternion function such that . The general two-sided quaternion linear canonical transform for the with respect to two linearly independent pure quaternions and can be expressed of the formandrespectively. 5.2. Convolution Theorem for General Two-Sided QLCT
In the classical domain [
28], a convolution operator is widely known as a signal processing algorithm in the theory of linear time-invariant systems. The output of any continuous time system is obtained by the convolution of the input signal with the system impulse response. The convolution is also important in designing the filters and in the reconstruction of the signals. According to these facts, the usefulness of convolution can be extended in a new domain. Furthermore, it is very interesting to consider the convolution definition and convolution theorem in the general two-sided QLCT domain. It shall be found that the convolution definition for the general two-sided QLCT is constructed by combining the quaternion convolution and the LCT convolution definitions [
29,
30].
Definition 7. For any two quaternion functions , we define the convolution and correlation operators related to the general two-sided QLCT asandrespectively. The above definition enables us to build the following important theorem which describes how the convolution of two quaternion-valued functions interacts with its general two-sided QLCT.
Theorem 5. Let be two quaternion-valued functions. If the decompositions of h and k are defined bythen convolution of h and k related to the general two-sided QLCT can be expressed as Proof. Using an inverse transform of the general two-sided QLCT (
41), we can rewrite the left-hand side of (
47) as
An application of two different splits to
h and
k defined by (
46) gives us
and thus
The proof is complete. □
From Theorem 5 we know that when
, the convolution theorem for the general two-sided QLCT above reduces to convolution theorem for general two-sided QFT [
19], that is,
6. Correlation Theorem for General Two-Sided QLCT
In the following we derive correlation theorem associated with the general two-sided QLCT. In this case we only apply two same splits to two quaternion function h and k as shown in the following theorem.
Theorem 6. Let be two quaternion-valued functions. If the decompositions of h and k are defined bythen the correlation of h and k related to general two-sided QLCT can be expressed as Proof. Taking into account inverse transform of the general two-sided QLCT (
41), we have
Subsequent calculation yields
The proof is complete. □
7. An Application
In this part, we aim to consider a simple application of the general two-sided QLCT for studying the generalized swept-frequency filters. The output of swept-frequency filters (compare to [
31,
32]) is then defined as
where
stands for quaternion convolution operator in the QFT domain. Assume that
is
real impulse response of the shift-invariant filter. By taking the general two-sided QLCT on both sides of (
53), we see that
If the matrix parameters are chosen as
and
, then the above identity can be expressed as
By changing the variables with
in the above expression we have
Consequently, we finally arrive at
8. Conclusions
In this work, based on the symmetric form of the two-sided quaternion Fourier transform, we have proved a variation on the Heisenberg-type uncertainty principle related to the two-sided quaternion Fourier transform. We also have introduced the general two-sided QLCT and provided its convolution definitions. In view of orthogonal plane split, we have derived convolution and correlation theorems related to general two-sided QLCT. We also have discussed its application to study the generalized swept-frequency filters.
Author Contributions
Conceptualization, M.B. and S.A.A.K.; Formal analysis, M.B. and S.A.A.K.; Funding acquisition, S.A.A.K.; Investigation, M.B. and S.A.A.K.; Methodology, M.B. and S.A.A.K.; Project administration, S.A.A.K.; Resources, M.B. and S.A.A.K.; Validation, M.B. and S.A.A.K.; Writing-original draft, M.B.; Writing-review and editing, M.B. and S.A.A.K. All authors have read and agreed to the published version of the manuscript.
Funding
Research Management Centre, Universiti Malaysia Sabah.
Institutional Review Board Statement
Not applicable.
Informed Consent Statement
Not applicable.
Data Availability Statement
Not applicable.
Acknowledgments
The first author is supported in part by Grant from Ministry of Education, Culture, Research and Technology, Indonesia under WCR scheme and the second author is fully supported by Universiti Malaysia Sabah, Malaysia. Special thank you to Research Management Centre and Faculty of Computing and Informatics, Universiti Malaysia Sabah.
Conflicts of Interest
The authors declare no conflict of interest.
References
- Hitzer, E. General two-sided quaternion Fourier transform, convolution and Mustard convolution. Adv. Appl. Clifford Algebr. 2017, 27, 381–395. [Google Scholar] [CrossRef]
- Hitzer, E.; Sangwine, S.J. The orthogonal 2D planes split of quaternions and steerable quaternion Fourier transformations. In Quaternion and Clifford Fourier Transforms and Wavelets, Trends in Mathematics; Hitzer, E., Sangwine, S.J., Eds.; Birkhäuser: Basel, Switzerland, 2013; pp. 15–40. [Google Scholar]
- Hitzer, E. Two-sided clifford with two square roots of -1 in Cl(p, q). Adv. Appl. Clifford Algebr. 2014, 24, 313–332. [Google Scholar] [CrossRef]
- Urynbassarova, D.; Li, B.-Z.; Zhang, Z.C. A convolution theorem for the polynomial Fourier transform. IAENG Int. J. Appl. Math. 2017, 47, 381–387. [Google Scholar]
- Xu, S.; Feng, L.; Chai, Y.; Hu, Y.; Huang, L. The properties of generalized offset linear canonical Hilbert transform and its applications. Int. J. Wavelets Multiresolut. Inf. Process. 2016, 14, 21. [Google Scholar] [CrossRef]
- Ahmed, H.M.; Elbarkouky, R.A.; Omar, O.A.M.; Ragusa, M.A. Models for covid-19 daily confirmed cases in different countries. Mathematics 2021, 9, 659. [Google Scholar] [CrossRef]
- Bahri, M.; Ashino, R. Logarithmic uncertainty principle for quaternion linear canonical transform. In Proceedings of the 2016 International Conference on Wavelet Analysis and Pattern Recognition, Jeju, Korea, 10–13 July 2016; pp. 140–145. [Google Scholar]
- Kou, K.I.; Ou, J.Y.; Morais, J. Uncertainty principles associated with quaternionic linear canonical transform. Math. Methods Appl. Sci. 2016, 39, 2722–2736. [Google Scholar] [CrossRef]
- Kou, K.I.; Ou, J.Y.; Morais, J. On uncertainty principle for quaternionic linear canonical transform. Abstr. Appl. Anal. 2013, 14, 725952. [Google Scholar] [CrossRef] [Green Version]
- Hu, X.X.; Kou, K.I. Quaternion Fourier and linear canonical inversion theorems. Math. Methods Appl. Sci. 2017, 40, 2421–2440. [Google Scholar] [CrossRef] [Green Version]
- Zhang, Y.N.; Li, B.-Z. Novel uncertainty principles for two-sided quaternion linear canonical transform. Adv. Appl. Clifford Algebr. 2018, 28, 15. [Google Scholar] [CrossRef]
- Yang, Y.; Dang, P.; Qian, T. Tighter uncertainty principles based on quaternion Fourier transform. Adv. Appl. Clifford Algebr. 2016, 26, 479–497. [Google Scholar] [CrossRef]
- Hitzer, E. Quaternion Fourier transform on quaternion fields and generalizations. Adv. Appl. Clifford Algebr. 2007, 20, 497–517. [Google Scholar] [CrossRef] [Green Version]
- De Bie, H.; De Schepper, N.; Ell, T.A.; Rubrecht, K.; Sangwine, S.J. Connecting spatial and frequency domains for the quaternion Fourier transform. Appl. Math. Comput. 2015, 271, 581–593. [Google Scholar] [CrossRef] [Green Version]
- Ell, T.A. Quaternion-fourier transforms for analysis of two-dimensional linear time-invariant partial differential systems. In Proceeding of the 32nd Conference on Decision and Control, IEEE Control Systems Society, San Antonio, TX, USA, 15–17 December 1993; pp. 1830–1841. [Google Scholar]
- Hitzer, E. Directional uncertainty principle for quaternion Fourier transform. Adv. Appl. Clifford Algebr. 2010, 20, 271–284. [Google Scholar] [CrossRef] [Green Version]
- Grigoryan, A.M.; Agaian, S. Tensor transform-based quaternion fourier transform algorithm. Inf. Sci. 2015, 320, 62–74. [Google Scholar] [CrossRef]
- Lian, P. Sharp Hausdorf-Young inequalities for the quaternion Fourier transforms. Proc. Am. Math. Soc. 2020, 148, 697–703. [Google Scholar] [CrossRef] [Green Version]
- Hitzer, E. Quaternionic Wiener-Khinchine theorems and spectral representation of convolution with steerable two-sided quaternion Fourier transform. Adv. Appl. Clifford Algebr. 2017, 27, 1313–1328. [Google Scholar] [CrossRef]
- Bahri, M.; Ashino, R. A convolution theorem related to quaternion linear canonical transform. Abstr. Appl. Anal. 2019, 2019, 3749387. [Google Scholar] [CrossRef]
- Li, Z.-W.; Gao, W.B.; Li, B.-Z. A new kind of convolution, correlation and product theorems related to quaternion linear canonical transform Signal. SIViP 2021, 15, 103–110. [Google Scholar] [CrossRef]
- Bülow, T. Hypercomplex Spectral Signal Representations for the Processing and Analysis of Images. Ph.D. Thesis, University of Kiel, Kiel, Germany, 1999. [Google Scholar]
- Morais, J.; Georgiev, S.; Sprosig, W. Real Quaternion Calculus Handbook; Birkhäuser: Basel, Switzerland, 2014. [Google Scholar]
- Jia, Z.G.; Wei, M.; Zhao, M.X.; Chen, Y. A new real structure-preserving quaternion QR algorithm. J. Comput. Appl. Math. 2018, 343, 26–46. [Google Scholar] [CrossRef] [Green Version]
- Chen, L.P.; Kou, K.I.; Liu, M.S. Pitt’s inequality and the uncertainty principle associated with the quaternion Fourier transform. J. Math. Anal. Appl. 2015, 423, 681–700. [Google Scholar] [CrossRef]
- Bahri, M.; Ashino, R. A variation on uncertainty principle and logarithmic uncertainty principle for continuous quaternion wavelet transforms. Abstr. Appl. Anal. 2017, 2017, 3795120. [Google Scholar] [CrossRef] [Green Version]
- Bahri, M. Simplex complex forms for two-sided quaternion linear canonical transform. J. Phys. Conf. Ser. 2019, 1180, 012013. [Google Scholar] [CrossRef]
- Debnath, L.; Shah, F.A. Wavelet Transforms and Their Applications; Birkhäuser: Basel, Switzerland, 2010. [Google Scholar]
- Wei, D.; Ren, Q.; Li, Y. A convolution and correlation theorem for the linear canonical transform and its application. Circuits Syst. Sig. Process. 2012, 31, 301–312. [Google Scholar] [CrossRef]
- Feng, Q.; Li, B.-Z. Convolution and correlation theorems for the two-dimensional linear canonical transform and its applications. IET Sig. Process. 2016, 10, 125–132. [Google Scholar] [CrossRef]
- Xiang, Q.; Qin, R. On the relationship between the linear canonicaltransform and the Fourier transform. In Proceedings of the 2011 4th International Congress on Image and Signal Processing, Shanghai, China, 15–17 October 2011; pp. 2214–2217. [Google Scholar]
- Bahri, M. Quaternion Linear Canonical Transform Application. Glob. J. Pure Appl. Math. 2015, 11, 8–24. [Google Scholar]
| Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
© 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).