Convolution Theorem for Canonical Cosine Transform and Their Properties
- 1. IOSR Journal of Mathematics (IOSR-JM)
e-ISSN: 2278-5728.Volume 6, Issue 1 (Mar. - Apr. 2013), PP 74-79
www.iosrjournals.org
www.iosrjournals.org 74 | Page
Convolution Theorem for Canonical Cosine Transform and Their
Properties
A. S. Gudadhe #
A.V. Joshi*
# Govt. Vidarbha Institute of Science and Humanities, Amravati. (M. S.)
* Shankarlal Khandelwal College, Akola - 444002 (M. S.)
Abstract: The Canonical Cosine transform, which is a generalization of the linear canonical transform, has
many applications in several areas, including signal processing and optics. In this paper we have introduced
convolution theorem, linearity property, derivative property, modulation property and Parseval’s identity for
the generalized canonical cosine transform.
Keywords: Fractional Fourier Transform, Linear Canonical Transform, Convolution.
I. Introduction:
As generalization of the Fourier Transform (FT), the Fractional Fourier Transform (FrFT) has been
used in several areas, including optics and signal processing. Many properties for this transform are already
known. In recent years, Almeida and Zayed derived product and convolution of two functions in a
usual manner and proved the convolution theorem in fractional Fourier transform domain. In the past decade,
FRFT has attracted much attention of the signal processing community, as the generalization of FT. The
relevant theory has been developed including uncertainty principle, sampling theory, convolution theorem.
A further generalization of FrFT is Linear Canonical Transform (LCT). Just as Fourier cosine
transform and Fourier sine transform are defined from Fourier Transform, similarly canonical cosine and
canonical sine transforms are defined from LCT by Pie and Ding [3]. We have discussed some properties of
Half canonical cosine transform in [2].
This paper emphasizes on defining generalized canonical cosine transform and deriving its convolution
theorem, then some properties of the canonical cosine transform are discussed and finally conclusions are given.
Notations and terminology as per [5]
II. Testing Function Space E :
An infinitely differentiable complex valued function on Rn
belongs to E (Rn
), if for each compact
set, SI where }0,,:{ tnRttS and for ,n
Rk
sup
( ) ( ) .
,
k
t D t
k t I
E
Note that space E is complete and a Frechet space, let E’ denotes the dual space of E.
2.1Definition: The generalized Canonical Cosine Transform ( )n
f RE can be defined by,
{CCT f (t)} (s) = < f(t), KC(t, s) > where,
t
b
s
t
b
ai
e
s
b
di
e
ib
st
C
K cos
2
2
2
2
.2
1
),(
Hence the generalized canonical cosine transform of a regular function ( )n
f RE can be defined by,
)1.1()(cos
2
1
)()(
22
22
dttfet
b
s
e
ib
stCCTf
t
b
ai
s
b
di
2.2 The Generalized Canonical Cosine Transform of Convolution
Now we introduced a special type of convolution and product for canonical cosine transform.
2.3 Definition: For any function , let us define the functions and by
)()(~,)(|)(|~
22
)(
2
)(
2
yvgeyvgyvgeyvg
yv
b
ai
yv
b
ai
and )()(
~
2
2
yfeyf
y
b
ai
- 2. Convolution Theorem For Canonical Cosine Transform And Their Properties
www.iosrjournals.org 75 | Page
For any two functions f and g, we define the Convolution operation by
(1.2)
Now we state and prove convolution theorem.
3.1 Convolution Theorem:
If and denote the Canonical Cosine transform of
respectively, then
Proof: From the definition of the Canonical Cosine transform, we have
dtdytgyft
b
s
y
b
s
ee
bi
ty
b
ai
s
b
d
i
)()(coscos2
2
1
0
)(
2
0
222
dtdytgyfty
b
s
ty
b
s
ee
bi
ty
b
ai
s
b
d
i
)()(})(cos)({cos
1
0
)(
2
0
222
dtdytgyfty
b
s
ee
bi
dtdytgyfty
b
s
ee
bi
ty
b
ai
s
b
d
i
ty
b
ai
s
b
d
i
)()()(cos
1
)()()(cos
1
0
)(
2
0
0
)(
2
0
222
222
21 II (1.3)
For I1, putting dvdtvty for limit when yvot , when vt
For I2, putting dvdtvty for limit when yvot , when vt
dvdyyvgyfv
b
s
eee
bi
dvdyyvgyfv
b
s
eee
bi
stgGsyfF
yv
yv
b
ai
y
b
ai
y
s
b
d
i
yv
yv
b
ai
y
b
ai
y
s
b
d
i
CC
)()()(cos
1
)()(cos
1
)))]((()[))]((([)3.1(
222
222
)(
22
0
)(
22
0
dvdyyvgyfv
b
s
eee
bi
dvdyyvgyfv
b
s
eee
bi
dvdyyvgyfv
b
s
eee
bi
dvdyyvgyfv
b
s
eee
bi
stgGsyfF
yv
yv
b
ai
y
b
ai
y
s
b
d
i
v
yv
b
ai
y
b
ai
y
s
b
d
i
yv
yv
b
ai
y
b
ai
y
s
b
d
i
v
yv
b
ai
y
b
ai
y
s
b
d
i
CC
)()(cos
1
)()(cos
1
)()(cos
1
)()(cos
1
)))]((()[))]((([
0 )(
22
0
0
)(
22
0
0 )(
22
0
0
)(
22
0
222
222
222
222
- 3. Convolution Theorem For Canonical Cosine Transform And Their Properties
www.iosrjournals.org 76 | Page
dvdyyvgyfv
b
s
eee
bi
dvdyyvgyfv
b
s
eee
bi
dvdyyvgyfv
b
s
eee
bi
dvdyyvgyfv
b
s
eee
bi
stgGsyfF
yv
yv
b
ai
y
b
ai
y
s
b
d
i
v
yv
b
ai
y
b
ai
y
s
b
d
i
y
v
yv
b
ai
y
b
ai
y
s
b
d
i
v
yv
b
ai
y
b
ai
y
s
b
d
i
CC
)()(cos
1
)()(cos
1
)()(cos
1
)()(cos
1
)))]((()[))]((([
0 )(
22
0
0
)(
22
0
0
)(
22
0
0
)(
22
0
222
222
222
222
4321)))]((()[))]((([ IIIIstgGsyfF CC (1.4) For
I4, putting dvdvvv for limit when yvyv , when 00 vv
dvdyvygyfv
b
s
eee
bi
dvdyvygyfv
b
s
eee
bi
I
y
v
vy
b
ai
y
b
ai
y
s
b
d
i
yv
vy
b
ai
y
b
ai
y
s
b
d
i
)()(cos
1
)()()()(cos
1
0
)(
22
0
0 )(
22
0
4
222
222
dvdyyvgyfv
b
s
eee
bi
dvdyyvgyfv
b
s
eee
bi
dvdyyvgyfv
b
s
eee
bi
dvdyyvgyfv
b
s
eee
bi
stgGsyfF
y
v
yv
b
ai
y
b
ai
y
s
b
d
i
v
yv
b
ai
y
b
ai
y
s
b
d
i
y
v
yv
b
ai
y
b
ai
y
s
b
d
i
v
yv
b
ai
y
b
ai
y
s
b
d
i
CC
)()(cos
1
)()(cos
1
)()(cos
1
)()(cos
1
)))]((()[))]((([)4.1(
0
)(
22
0
0
)(
22
0
0
)(
22
0
0
)(
22
0
222
222
222
222
dvdyyvgyfv
b
s
eee
bi
dvdyyvgyfv
b
s
eee
bi
stgGsyfF
v
yv
b
ai
y
b
ai
y
s
b
d
i
v
yv
b
ai
y
b
ai
y
s
b
d
i
CC
)()(cos
1
)()(cos
1
)))]((()[))]((([
0
)(
22
0
0
)(
22
0
222
222
dvev
b
s
dyyvgvygee
bibi
v
b
ai
v
b
ai
s
b
d
i
yf
222
2
0 0
2
cos|))(|~)(~(
2
2
1
)(
~
by (1.2)
dvvgfv
b
s
ee
biib
ee v
b
ai
s
b
div
b
ai
s
b
di
)})(~*
~
{(cos
2
2 0
22
22 22
22
by (2.3)
)()(~*
~
2
)))]((()[))]((([
22
22
svgfF
ib
ee
stgGsyfF C
v
b
ai
s
b
di
CC
- 4. Convolution Theorem For Canonical Cosine Transform And Their Properties
www.iosrjournals.org 77 | Page
4.1 Some operational results:
4.1.1 Linearity property of canonical cosine transformations:
If {CCT f(t)}(s), {CCT g(t)}(s) denotes generalized canonical cosine transform of f(t), g(t) and P1, P2
are constants then )())(()()(()())()(( 2121 stgCCTPstfCCTPstgPtfPCCT
Proof is simple and hence omitted.
4.1.2 Derivative (with respect to parameter) of canonical cosine transform:
If {CCT f(t)}(s) denotes generalized canonical cosine transform, then,
))]}((.[{
1
))}(({.))}(({ stftCST
b
stCCTf
b
d
sistCCTf
ds
d
Proof: We have,
dttft
b
s
ee
ibds
d
stfCCT
ds
d t
b
ai
s
b
di
)(cos
2
1
))}(({
22
22
dttft
b
s
e
s
e
ib
stfCCT
ds
d s
b
di
t
b
ai
)(cos.
2
1
))}(({
22
22
dttft
b
s
es
b
d
it
b
s
e
b
t
e
ib
s
b
di
s
b
di
t
b
ai
)(cos..sin.
2
1
222
222
dttft
b
s
ee
ibb
d
sdttftt
b
s
e
b
e
ib
ii
s
b
di
t
b
ai
s
b
di
t
b
ai
)(cos
2
1
.)](.[sin.
1
2
1
)(
2222
2222
))}(({.))]}((.[[
1
{ stCCTf
b
d
sstftCST
b
i
))]}((.[{
1
))}(({.))}(({ stftCST
b
stCCTf
b
d
sistCCTf
ds
d
4.1.3 Modulation property of canonical cosine transform:
If {CCT f(t)} (s) denotes generalized canonical cosine transform of f(t) then,
)()()()(
2
)()(.cos
2
2
b
bzs
etfCCT
b
bzs
etfCCT
e
stfztCCT idszidsz
dbz
i
Proof: By definition of CCT,
dttftz
b
s
tz
b
s
ee
ib
stfztTCC
t
b
ai
s
b
di
)(coscos.
2
1
2
1
))}((.cos{
22
22
})(..cos
2
1
)(.cos
2
1
{
2
1
))}((.cos{
222
222
2
)(
22
2
)(
22
dttftz
b
s
eeee
ib
dttftz
b
s
eeee
ib
stfztTCC
zdb
i
idsz
bzs
b
di
t
b
ai
zdb
i
idsz
bzs
b
di
t
b
ai
})(..
)(
cos
2
1
)(.
)(
cos
2
1
{
2
))}((.cos{
22
22
2
)(
22
)(
22
2
dttfet
b
bzs
ee
ib
dttfet
b
bzs
ee
ib
e
stfztTCC
idsz
bzs
b
di
t
b
ai
idsz
bzs
b
di
t
b
aizdb
i
- 5. Convolution Theorem For Canonical Cosine Transform And Their Properties
www.iosrjournals.org 78 | Page
)()()()(
2
)()(.cos
2
2
b
bzs
etfCCT
b
bzs
etfCCT
e
stfztCCT idszidsz
dbz
i
4.1.4 If {CCT f(t)} (s) denotes generalized canonical cosine transform of f(t) then,
)()()()(
2
)()()(.sin
2
2
b
bzs
etfCCT
b
bzs
etfCST
e
istfztCCT idszidsz
zdb
i
Proof: By definition of CCT,
dttfztt
b
s
ee
ib
stfztTCC
t
b
ai
s
b
di
)(sincos.
2
1
))}((.sin{
22
22
dttftz
b
s
tz
b
s
ee
ib
stfztTCC
t
b
ai
s
b
di
)(sinsin.
2
1
2
1
))}((.sin{
22
22
})(..sin
2
1
)(.sin
2
1
{
2
1
))}((.sin{
222
222
2
)(
22
2
)(
22
dttftz
b
s
eeee
ib
dttftz
b
s
eeee
ib
stfztTCC
zdb
i
idsz
bzs
b
di
t
b
ai
zdb
i
idsz
bzs
b
di
t
b
ai
})(..cos
2
1
)(
)(.sin
2
1
){(
2
)(
))}((.sin{
22
22
2
)(
22
)(
22
2
dttfet
b
bzs
ee
ib
i
dttfet
b
bzs
ee
ib
i
ei
stfztTCC
idsz
bzs
b
di
t
b
ai
idsz
bzs
b
di
t
b
aizdb
i
)()()()(
2
)()()(.sin
2
2
b
bzs
etfCCT
b
bzs
etfCST
e
istfztCCT idszidsz
zdb
i
5.1 Parseval’s Identity for canonical cosine transform:
If )(tf and )(tg are the inversion canonical cosine transform of )(sFC and )(sGC respectively, then
(1) dssGsFdttgtf CC )(.)(2)(.)(
and (2) dssFdttf C
22
)(2)(.
Proof: By definition of CCT,
dttgt
b
s
ee
ib
stgCCT
t
b
ai
s
b
di
)(cos
2
1
)()(
22
22
---------------- by (1.1)
Using the inversion formula of CCT
dssGt
b
s
ee
b
i
tg C
s
b
di
t
b
ai
)(cos
2
)(
22
22
Taking complex conjugate we get,
dssGt
b
s
ee
b
i
tg C
s
b
di
t
b
ai
)(cos
2
)(
22
22
- 6. Convolution Theorem For Canonical Cosine Transform And Their Properties
www.iosrjournals.org 79 | Page
dssGt
b
s
ee
b
i
dttfdttgtf C
s
b
di
t
b
ai
)(cos
2
)()(.)(
22
22
Changing the order of integration, we get,
dttft
b
s
ee
ib
ib
dssG
b
i
dttgtf
s
b
di
t
b
ai
C )(cos
2
1
2
1
1
)(
2
)(.)(
22
22
(1.5)
(2) Now putting )()( tgtf in equation (1.5), we get
dssFdttf C
22
)(2)(
Table for canonical cosine transform
S.N
.
f(t) FC(s)
1 ))()(( 21 tgPtfP )())(()()(( 21 stgCCTPstfCCTP
2
)(.cos tfzt
)()()()(
2
2
2
b
bzs
etfCCT
b
bzs
etfCCT
e idszidsz
zdb
i
3
)(.sin tfzt
)()()()(
2
)(
2
2
b
bzs
etfCCT
b
bzs
etfCST
e
i idszidsz
zdb
i
III. Conclusion:
The Convolution of generalized canonical cosine transform is developed in this paper. Operation
transform formulae proved in this paper can be used, when this transform is used to solve ordinary or partial
differential equation.
References:
[1] Almeida L B (1997): Product and convolution theorems for the fractional Fourier transform. IEEE Signal Processing Letters,4(1):
Page No.15—17
[2] Gudadhe A. S., Joshi A.V. (2013): On Generalized Half Canonical Cosine Transform, IOSR-JM Volume X, Issue X.
[3] Pie S. C., Ding J. J.( 2002): Fractional cosine, sine and Hartley transform; trans. On signal processing, Vo. 50, No. 7, P. 1661-1680,
pub.
[4] Zayed A. I., (1998): A convolution and product theorem for the fractional Fourier transform, IEEE Sig. Proc. Letters, Vol. 5, No. 4,
101-103.
[5] Zemanian A. H., (1965): Distribution Theory and Transform Analysis, Mc. Graw Hill, New York.