Properties of Quaternion Fourier Transforms
|
|
- Louisa Lynn Carter
- 6 years ago
- Views:
Transcription
1 Properties of Quaternion Fourier Transforms Dong Cheng and Kit Ian Kou Department of Mathematics, Faculty of Science and Technology, University of Macau, Macao, China arxiv: v [math.ca] 8 Jun 06 Abstract With the recent popularity of quaternion and quaternion Fourier transforms (QFTs technique in physics and engineering applications, there tends to be an inordinate degree of interest placed on the properties of QFTs. Due to the noncommutativity of multiplication of quaternions, there are different types of QFTs and we focus on the right-sided QFT (QFT and two-sided QFT (SQFT. In this paper, we begin by considering the behavior of the QFT on the space L (,H. Theenergy-preservedpropertydeducestotheextension ofqftonl (,H. We are largely exploiting the relation between QFT and SQFT and some elementary properties of QFT can be transformed to SQFT. As a generalization of QFTs, the properties of quaternion linear canonical transforms (QLCTs are also studied on the basis of existing properties of QFTs. Keywords: quaternion Fourier transforms; quaternion linear canonical transforms; inversion theorem; Plancherel theorem 000 Mathematics Subject Classification: 4A38, 4B0, 43A3, 43A50 Introduction The quaternion Fourier transforms (QFTs play a vital role in the representation of (hypercomplex signals. They transform real (or quaternion-valued D signals into a quaternion-valued frequency domain signals. The four components of the QFT separate four cases of symmetry into real signals instead of only two as in the complex Fourier Transform (FT. In [6, 8] authors used the QFT to proceed color image analysis. The study [, 7] implemented the QFT to design a color image digital watermarking scheme. chengdong70@63.com kikou@umac.mo
2 esearchers in [3] applied the QFT to image pre-processing and neural computing techniques for speech recognition. ecently, certain asymptotic properties of the QFT were analyzed and a straightforward generalization of the classical Bochner-Minlos theorem to the framework of quaternion analysis was derived []. In classical Fourier theory, if f is integrable (L (,C, then the Fourier transform f(ξ is well-deمحned by f(ξ = f(te iξt dt. (. Moreover, if f is also integrable, then f(t = f(ξe iξt dξ for almost every t. The integral (. deمحning the Fourier transform is not suitable for square integrable function (L (,C. But There is a natural and elegant theory of Fourier transform for square integrable functions. In fact, the Fourier transform F on L (,C is not only isometric but also onto. This desirable property is very important in signal processing. One may naturally asks what are the analogous results for QFTs of quaternion-valued signals? The previous contributions on inversion theorem and energy-preserved property of QFTs were developed in [6, ]. On one hand, however, the existing results are not well established systematically. On the other hand, the prerequisites for setting up of the established theorem did not be studied completely. Therefore it is of great interest to progress with a function theory of QFT. In light of this, the inversion theorem on L (,H and Plancherel theorem of QFTs are investigated thoroughly in this paper. Moreover, as a generalization of QFTs, the properties of quaternion linear canonical transforms (QLCTs are also developed combining with the existing properties of QFTs. The rest of the paper is organized as follows. In the next section, we recall some basic knowledge of quaternion algebra. Section 3 investigates the inversion theorem on L (,H and Plancherel theorem of right-sided QFT. In section 4, we establish the relation between QFT and SQFT and study the elementary properties of SQFT. In section 5, the properties of quaternion linear canonical transforms (QLCTs are studied on the basis of existing properties of QFTs. Preliminaries. Quaternion algebra Throughout the paper, let H := {q = q 0 +iq +jq +kq 3 q 0,q,q,q 3 }, be the Hamiltonian skew field of quaternions, where the elements i, j and k obey the Hamiltonلاs multiplication rules: ij = ji = k, jk = kj = i, ki = ik = j, i = j = ijk =.
3 For every quaternion q = q 0 + q, q = iq + jq + kq 3, the scalar and vector parts of q, are deمحned as Sc(q = q 0 and Vec(q = q, respectively. If q = Vec(q, then q is called pure imaginary quaternion. The quaternion conjugate is deمحned by q = q 0 q = q 0 iq jq kq 3, and the norm q of q deمحned as q = qq = qq = m=3 m=0 q m. Then we have q = q, p+q = p+q, pq = q p, pq = p q, p,q H. Using the conjugate and norm of q, one can deمحne the inverse of q H\{0} by q = q/ q. The multiplication of two quaternions is noncommutative, but Sc(pq = Sc(qp p,q H. (. The quaternion exponential function e q is deمحned by means of an inمحnite series as e q := Analogous to the complex case one may derive a closed-form representation: n=0 q n n!. e q = e q 0 (cos q + q q sin q. The quaternion has subsets C µ := {a+bµ : a,b, µ =,µ = Vec(µ}. For each complex. xedمح unit pure imaginary quaternion µ, C µ is isomorphic to the. Quaternion module L p (,H The left quaternion module L p (,H(p =, consists of all H-valued functions whose pth power is Lebesgue integrable on : } L p (,H := {f f : H, f plp(,h := f(x,x p dx dx <. X Like the complex case, the left quaternion module L (,H consists of all essentially bounded measurable H-valued functions on. The left quaternionic inner product of f,g L (,H is deمحned as < f,g > L (,H= f(x,x g(x,x dx dx. This inner product leads to a scalar norm f =< f,f > L (X,H which coincides with the usual L -norm for f, considered as a vector-valued function. In fact, L (,H is a left quaternionic Hilbert space with inner product <, > L (,H (see [5]. For simplicity of notation, we use <, > in place of <, > L (,H in the following. It is not difficult to verify that < pf,qg >= p < f,g > q (. 3
4 In fact, f = f 0 + if + jf + kf 3 L p (,H(p =,, if and only if f m (m = 0,,,3 L p (,. Then we can consider every f L p (,H(p =,, as a linear combination of real (complex L p functions. Therefore, some properties of real (complex L p functions can be natural extended to quaternionic L p functions. The corresponding complex versions of following results can be found in [5, 7]. Proposition. If {f n } is a cauchy sequence in L p (,H(p =,,, with limit f, then {f n } has a subsequence which converges pointwise almost every (x,x to f. Proposition. Let p ǫ (x,x = ǫ be the Poisson kernel, then π (ǫ +x (ǫ +x. lim ǫ 0 f p ǫ f p = 0 for every f L p (,H(p =,;. lim ǫ 0 (f p ǫ (x,x = f(x,x, if f(x,x L (,H is continuous at a point (x,x. 3 The right-sided quaternion Fourier transform The quaternion Fourier transform (QFT was rstمح deمحned by Ell to analyze linear timeinvariant systems of partial differential equations [9]. Later, some constructive works related to the QFT and its application to color image processing were studied (see [6, 4, 6,, 8, ]. The non-commutativity of the quaternion multiplication leads to different types of quaternion Fourier transformations (see [0]. In this section, we consider the right-sided quaternion Fourier transform. 3. Theright-sided quaternion Fourier transform pairs in L (,H Definition 3. (QFT For every f L (,H, the factored, right-sided quaternion Fourier transform of f is defined by (F r f(ω,ω := f(x,x e iω x dx dx. π The following results are related to our study of the inversion theorem and Plancherel theorem of QFT. We use the integral representations to express the convolutions. Proposition 3. If f L (,H, put f(x,x := f( x, x and define g(x,x := ( f f(x,x = f(y,y f(x +y,x +y dy dy, then (f p ǫ (x,x = P(ǫω,ǫω (F r f(ω,ω e jω x e iω x dω dω π and Sc((g p ǫ (0,0 = P(ǫω,ǫω (F r f(ω,ω dω dω, where P(ω,ω = e ω ω. 4
5 Proof. Since p ǫ (x y,x y = π ǫ e ǫω e iω y e iω x dω ǫ +(x y = e ǫω e iω ǫ y x dω π ǫ +(x y eiω = e ǫω e iω y e ǫω e jω (x y dω 4π e iω x dω = P(ǫω 4π,ǫω e iω y e jω y e jω x e iω x dω dω, then Then (f p ǫ (x,x = dy dy f(y,y P(ǫω 4π,ǫω e iω y e jω y e jω x e iω x dω dω = P(ǫω 4π,ǫω dω dω f(y,y e iω y e jω y e jω x e iω x dy dy = P(ǫω,ǫω (F r f(ω,ω e jω x e iω x dω dω. π Note that g(x,x = ( f f(x,x = f( y, y f(x y,x y dy dy = f(y,y f(x +y,x +y dy dy. Sc((g p ǫ (0,0 ( = Sc g(y,y p ǫ ( y, y dy dy ( [ = Sc f(s,s f(s +y,s +y ds ds ]p ǫ ( y, y dy dy ( = Sc ds ds f(s,s f(s +y,s +y p ǫ ( y, y dy dy ( = Sc ds ds f(s,s f(z,z p ǫ (s z,s z dz dz ( = Sc ds 4π ds f(s,s f(z,z dz dz P(ǫω,ǫω e iω z e jω z e jω s e iω s dω dω ( 6 = Sc dω 4π dω P(ǫω,ǫω ds ds e jω s e iω s f(s,s f(z,z e iω z e jω z dz dz 6 = P(ǫω,ǫω (F r f(ω,ω dω dω 5
6 which completes the proof. To لاauthors knowledge, the inversions of QFT have not been investigated for absolutely integrable quaternion-valued function thoroughly. By Proposition.,. and 3., we give an inversion theorem about QFT as follows. Theorem 3.3 (Inversion of QFT If f,f r f L (,H and g(x,x = (F r f(ω,ω e jω x e iω x dω dω, π then f(x,x = g(x,x for almost every (x,x. Proof. Form Proposition 3., (f p ǫ (x,x = π P(ǫω,ǫω (F r f(ω,ω e jω x e iω x dω dω. (3. The integrands on the right side of (3. are bounded by (F r f(ω,ω, and since P(ǫω,ǫω as ǫ 0, then π P(ǫω,ǫω (F r f(ω,ω e jω x e iω x π (F rf(ω,ω e jω x e iω x as ǫ 0. Hence, the right side of (3. converges to g(x,x, for every (x,x, by the dominated convergence theorem. From Proposition. and., we see that that there is a sequence {ǫ n } such that ǫ n 0 and lim n (f p ǫn (x,x = f(x,x, for almost every (x,x. Hence f(x,x = g(x,x for almost every (x,x. If f L (,H, Theorem 3.3 indicates that F r f = 0 for almost every (ω,ω implies f = 0 for almost every (x,x. Corollary 3.4 (Uniqueness of QFT If f,g L (,H and(f r f(ω,ω = (F r g(ω,ω for almost every (ω,ω, then f(x,x = g(x,x for almost every (x,x. Now we see that, under suitable conditions, the original signal f can be reconstructed from F r f by the inverse right-sided quaternion Fourier transform (IQFT. Definition 3.5 (IQFT For every F r L (,H, the inverse right-sided quaternion Fourier transform of F r is defined by (Fr F(x,x := F r (ω,ω e jω x e iω x dω dω. π emark 3.6 Both F r and Fr are bounded linear transformation from L (,H into L (,H. We will see later, both F r L L and F r L L can be extended to L (,H. As an operator on L (,H, Fr is the inversion of F r. 6
7 Next we present some properties of inverse QFT. Proposition 3.7 Suppose that F r L (,H.. If G r (ω,ω = F r (ω,ω iω L (,H, then the partial derivative of F with respect to x exists and (F r F r (x, x = (Fr G r (x,x. x r F r. If G r (ω,ω = F r (ω,ω jω L (,H, then the partial derivative of F r F r with respect to x exists and (F F r x (x,x = (F r G r (x,x. 3. If G 3r (ω,ω = F r (ω,ω kω ω L (,H, then (F r F r x x (x, x = (F r G 3r (x,x. An important result called the multiplication formula in classical Fourier analysis can be generalized to QFT. Before to give the formula, we introduce an auxiliary transform for f(x,x = f 0 (x,x + if (x,x + jf (x,x + kf 3 (x,x which is deمحned by αf(x,x := f 0 (x,x +if (x, x +jf ( x,x +kf 3 ( x, x. Then we obtain the following result. Proposition 3.8 Suppose that f,g L (,H, h := αg, H r := F r h and F r := F r f, then F r (x,x g(x,x dx dx = f(x,x H r (x,x dx dx. (3. Moreover, if g is contained in L (,H, we have g = h. Proof. Write g = g 0 +ig +jg +kg 3, then e iω x g(ω,ω dω dω = g 0 (ω,ω e iω x dω dω + ig (ω,ω e iω x e jω x dω dω + jg (ω,ω e iω x dω dω + kg 3 (ω,ω e iω x e jω x dω dω = g 0 (ω,ω e iω x dω dω + ig (ω, ω e iω x dω dω + jg ( ω,ω e iω x dω dω + kg 3 ( ω, ω e iω x dω dω = h(ω,ω e iω x dω dω = H r (x,x. 7
8 Applying Fubiniلاs theorem, we have F r (ω,ω g(ω,ω dω dω ( = f(x,x e iω x dx dx g(ω,ω dω dω ( = f(x,x e iω x g(ω,ω dω dω dx dx = f(x,x H r (x,x dx dx. If g L (,H, it is easy to verify that g = h by deمحnition of α. 3. The Plancherel theorem of QFT In the complex case, the Plancherel theorem indicates that if f L L, and it turns out that f L and f = f, where f is the classical Fourier transform of f. Moreover, this isometry of L L into L extends to an isometry of L onto L, and this extension deمحnes the Fourier transform of every f L. The convolution theorem plays a vital role in proving the Plancherel theorem (see [5, 7]. However, the classical convolution theorem no longer holds for the QFT. The Plancherel theorem of QFT was discussed in recent research papers (see [6],[],[]. We give a restatement of Plancherel theorem here since the prerequisites for setting up of the theorem may not be put forward so clearly in recent research papers. It is probably worth pointing out that Proposition 3. plays a key role in our proof. Theorem 3.9 If f L (,H L (,H, then F r f L (,H and Parseval s identity F r f = f holds. Proof. We xمح f L (,H L (,H, put f(x,x := f( x, x and deمحne g(x,x := ( f f(x,x =< f,f x, x >= f(y,y f(x +y,x +y dy dy, where f x, x (y,y denotes a translate of f. Since (x,x f x, x is a continuous mapping of into L (,H and noticing that the continuity of the inner product, we see that g(x,x is a continuous function. By Cauchy-Schwarz inequality, g(x,x f x, x f = f. Thus g is bounded. Furthermore, g L (,H since f L (,H and f L (,H. Since g is continuous and bounded, Proposition. shows that limsc((g p ǫ (0,0 = Sc(g(0,0 = f ǫ 0. 8
9 On the other hand, since g L (,H, by Proposition 3.: Sc((g p ǫ (0,0 = P(ǫω,ǫω (F r f(ω,ω dω dω. Since0 P(ǫω,ǫω (F r f(ω,ω increases to (F r f(ω,ω asǫ 0, themonotone convergence theorem gives limsc((g p ǫ (0,0 = lim P(ǫω,ǫω (F r f(ω,ω dω dω ǫ 0 = limp(ǫω,ǫω (F r f(ω,ω dω dω ǫ 0 = (F r f(ω,ω dω dω = F r f. ǫ 0 Therefore, F r f L (,H and F r f = f. From Theorem 3.9, F r L L is an isometry of L (,H L (,H into L (,H. Since L (,H L (,H is a dense subset of L (,H. Therefore, there exists a unique bounded (continuous extension, Ψ r, of F r L L to all of L (,H. If f L (,H, F r = Ψ r f is deمحned by the L limit of the sequence {F r f k }, where {f k } is any sequence in L (,H L (,H converging to f in the L norm. If we choose f k (x,x = f(x,x χ [ k,k] (x,x, we have F r (ω,ω = l.i.m. k f(x,x e ix ω e jx ω dx dx. π [ k,k] where f = l.i.m. k f k means f f k 0 as k. We call F r = Ψ r f is the QFT of f on L (,H. The multiplication formula (3. easily extends to L (,H. The left H-linear operator Ψ r on L (,H is an isometry. So Ψ r is a one-to-one mapping. Moreover, we can show that Ψ r is onto. Theorem 3.0 The QFT Ψ r is a unitary operator on L (,H. Proof. Firstly, we show that the range of Ψ r denoted by (Ψ r is a closed subspace of L (,H. Let {F k } be a sequence in (Ψ r converging to F r in L norm sense. we now prove that F r (Ψ r. Suppose that Ψ r f k = F k. Then the isometric property shows that Ψ r is continuous and {f k } is also a Cauchy sequence. The completeness of L (,H implies that {f k } converges to some f L (,H, and the continuity of Ψ r shows that Ψ r f = l.i.m. k Ψ rf k = F r. If (Ψ r were not all of L (,H, as every closed subspace of Hilbert space L (,H has an orthogonal complement, we could ndمح a function u such that F r (x,x u(x,x dx dx = 0 9
10 for all f L (,H and u 0. Let g = u,h = αg, by multiplication formula, f(x,x H r (x,x dx dx = F r (x,x g(x,x dx dx = 0 for all f L (,H. Pick f = H r, this implies that H r (x,x = 0 for almost every (x,x, contradicting the fact that H r = h = g = u 0. Next result shows that the mapping Ψ r is a Hilbert space isomorphism of L (,H, that is, preserving inner product. Theorem 3. Let f,g L (,H and F r = Ψ r f,g r = Ψ r g. Then f(x,x g(x,x dx dx = F r (ω,ω G r (ω,ω dω dω. Proof. Let p 0 +ip +jp +kp 3 = f(x,x g(x,x dx dx and q 0 +iq +jq +kq 3 = F r (ω,ω G r (ω,ω dω dω. From the Parsevalلاs identity, we have f +g = f + g +p 0 = F r +G r = F r + G r +q 0. Thus p 0 = q 0. By using (. and applying Parsevalلاs identity to f+ig = F r+ig r, f +jg = F r +jg r, f +kg = F r +kg r respectively, we can get p m = q m, (m =,,3 which completes the proof. Theorem 3. The inverse f = Ψ r F r is the L limit of the sequence {Fr F rk } k, where {F rk } k is any sequence in L (,H L (,H converging to F r in the L norm. If we choose F rk = F r χ [ k,k], we have f(x,x = l.i.m. k F r (ω,ω e jx ω e ix ω dω dω. π [ k,k] In particular, if F r L (,H L (,H, then f(x,x = F r (ω,ω e jx ω e ix ω dω dω. π 0
11 Proof. For Ψ r B(L (,H, the quaternionic iesz representation theorem (see [5] guarantees that there exists a unique operator Ψ r B(L (,H, which is called the adjoint of Ψ r, such that for all f,g L (,H, (Ψ r f,g = (f,ψ r g. Since Ψ r is unitary, then Ψ r = Ψ r. For any xedمح F r L (,H, let {F rk } k be an arbitrary sequence in L (,H L (,H converging to F r in the L norm, we have < g,ψ r F r > =< G r,f r > = lim < G r,f rk > k ( = lim k ( = lim g(x,x k = lim < g,fr F rk >=< g,l.i.m. k k F r F rk > g(x,x e ix ω e jx ω dx dx F rk (ω,ω e jx ω e ix ω dω dω F rk (ω,ω dω dω dx dx for all g L (,H L (,H. Thus f = Ψ r F = Ψ rf r = l.i.m. k F r F rk. In particular, if F r L (,H L (,H, then which completes the proof. f(x,x = π F r (ω,ω e jx ω e ix ω dω dω emark 3.3 Since Ψ r (Ψ r coincides with F r (F r in L (,H L (,H. For simplicity of notations, in the following, by capital letter F r, we mean the QFT of f L (,H L (,H if no otherwise specified. 4 The two-sided quaternion Fourier transform In this section, we study the two-sided (sandwich quaternion Fourier transform (SQFT. 4. Thetwo-sided quaternion Fouriertransform pairs in L (,H Definition 4. (SQFT For every f L (,H, the factored, two-sided quaternion Fourier transform of f is defined as (F s f(ω,ω := e iω x f(x,x dx dx. π Unlike the QFT, SQFT is not a left H-linear operator. But SQFT is left C i -linear and right C j -linear. Moreover, SQFT could establish relationship with QFT through the following transform.
12 Definition 4. Iff(x,x = f 0 (x,x +if (x,x +jf (x,x +kf 3 (x,x L p (X,H(p =,, the transform β is defined by βf(x,x := f 0 (x,x +if (x,x +jf ( x,x +kf 3 ( x,x. Proposition 4.3 Suppose that f,g L p (X,H(p =,. Then the following assertions hold.. β is a left C i -linear bijection mapping on L p (X,H. So the inverse of β can be defined and β is actually equal to β itself.. If f L (,H then F s f = F r (βf. (4. Moreover, if f is C i -valued or f is even with respect to first variable, then F s f = F r f. 3. If f,g L (,H then Sc(< βf,βg > = Sc(< f,g >, Sc(i < βf,βg > = Sc(i < f,g >. In particular f = βf. Proof. The assertion ( is a direct consequence of the deمحnition of β. To prove (, write f in form of f 0 +if +jf +kf 3 and let h := βf. Then F s f(ω,ω = e iω x f(x,x dx dx = f 0 (x,x e iω x dx dx + if (x,x e iω x dx dx + jf (x,x e iω x dx dx + kf 3 (x,x e iω x ddx dx = f 0 (x,x e iω x dx dx + if (x,x e iω x dx dx + jf ( x,x e iω x dx dx + kf 3 ( x,x e iω x ddx dx = h(x,x e iω x dx dx = F r h(ω,ω. If f is C i -valued or f is even with respect to rstمح variable, then βf = f. It follows that F s f = F r (βf.
13 Now we prove (iii. If f,g L (,H, then Sc(< βf,βg > = f 0 (x,x g 0 (x,x dx dx + f (x,x g (x,x dx dx + f ( x,x g ( x,x dx dx + f 3 ( x,x g 3 ( x,x dx dx = f 0 (x,x g 0 (x,x dx dx + f (x,x g (x,x dx dx + f (x,x g (x,x dx dx + f 3 (x,x g 3 (x,x dx dx = Sc(< f,g >. Since β is left C i -linear, then Sc(i < f,g > = Sc(< if,g > = Sc(< β(if,g > = Sc(< iβf,g > = Sc(i < βf,βg >. At last which completes the proof. βf = f 0 + f + f + f 3 = f Theorem 4.4 (Inversion of SQFT If f,f s f L (,H and g(x,x = e iω x (F s f(ω,ω e jω x dω dω, (4. π then f(x,x = g(x,x for almost every (x,x. Proof. By invoking (ii of Proposition 4.3 we have F s f = F r (βf. Let h := Fr (F s f, then (βf(x,x = h(x,x for almost every (x,x by Theorem 3.3. To prove f(x,x = g(x,x for almost every (x,x, it suffices to verify βg = h. Note that h(x,x = (F s f(ω,ω e jω x e iω x dω dω. π It is easy to see that Sc(h(x,x = Sc(g(x,x by invoking equation (.. Similarly we have Sc(h(x,x i = Sc(g(x,x i. Since Then h(x,x j = π (F s f(ω,ω e jω x je iω x dω dω. ( Sc(h(x,x j = Sc (F s f(ω,ω e jω x je iω x dω dω π ( = Sc e iω x (F s f(ω,ω e jω x jdω dω π = Sc(g( x,x j. 3
14 Analogously, we have Sc(h(x,x k = Sc(g( x,x k. Then we conclude that βg = h which completes the proof. So we can deمحne the inverse two-sided quaternion Fourier transform by equation (4. or equivalently by β F r F s. Definition 4.5 (ISQFT For every F s L (,H, the inverse two-sided quaternion Fourier transform of F s is defined by (Fs F s (x,x := e iω x F s (ω,ω e jω x dω dω. π 4. The Plancherel theorem of SQFT In section 3., we extend F r L L to L (,H. The QFT on L (,H has more symmetry than QFT in L (,H. The relation F s f = F r (βf drives us to extend F s L L to L (,H. Definition 4.6 For every f L (,H, the SQFT of Ψ s f is defined by Ψ s f := Ψ r (βf (4.3 In fact, we can deمحne Ψ s starting from the original deمحnition 4. by taking L norm limit. (4.3 gives us a different but actually equivalent form of Ψ s. Theorem 4.7 Suppose that f,g s L (,H, Then the following assertions hold.. Ψ s f defined by (4.3 is equal to the L limit of the sequence {F s f k } k, where {f k } k is any sequence in L (,H L (,H converging to f in the L norm. If f L (,H L (,H, then Ψ s f = F s f.. Ψ s is a bijection on L (,H and Ψ s G s = β Ψ r G s. Furthermore, Ψ s G s is equal to the L limit of the sequence {Fs G sk } k, where {G sk } k is any sequence in L (,H L (,H converging to G s in the L norm. If G s L (,H L (,H, then Ψ s G s = Fs G s. Proof. The assertion (i is a consequence of (4.3 and deمحnition of Ψ r. The assertion (ii is a consequence of (4.3 and Theorem 3.. As an immediate consequence of Theorem 3. and Theorem 3., we present the following result. Theorem 4.8 If f,g L (,H, then < Ψ s f,g >=< βf,ψ r g >. HavingdeمحnedtheSQFTforfunctionsinL (,H, weobtainthefollowingparsevalلاs identity. 4
15 Theorem 4.9 Suppose that f,g L (,H, F s = Ψ s f,g s = Ψ s g. Let p 0 +ip +jp +kp 3 = f(x,x g(x,x dx dx and q 0 +iq +jq +kq 3 = F s (ω,ω G s (ω,ω dω dω. Then F s = f and p m = q m, (m = 0,. Moreover, if both f and g are C i -valued or even with respect to first variable, then p m = q m, (m = 0,,,3. Proof. Firstly, we show that Parsevalلاs identity of SQFT holds. Applying Parsevalلاs identity of QFT and (iii of Proposition 4.3, we have F s =< Ψ sf,ψ s f > =< Ψ r (βf,ψ r (βf > =< βf,βf > = βf = f. Byinvoking Parsevalلاs identity ofsqftto f+g = F s+g s, f+ig = F s+ig s respectively, we get p m = q m, (m = 0,. If both f and g are C i -valued or even with respect to rstمح variable, we have Ψ s f = Ψ r f and Ψ s g = Ψ r g, therefore < Ψ s f,ψ s g >=< Ψ r f,ψ r g >=< f,g >, that is p m = q m, (m = 0,,,3. emark 4.0 By applying (. Hitzer [] proved p 0 = q 0, it follows that F s = f. We have shown that p is also equal to q. One may be wondering what is the relationship between p and q (p 3 and q 3? At the moment, we do not know their relation. Since Ψ s is not left H-linear, so f + jg F s + jg s. In fact, f + gj = F s + G s j. It follows that Sc(< f,gj > = Sc(< F s,g s j > rather than Sc(< f,jg > = Sc(< F s,jg s >. However, Sc(< f,jg > = Sc(< F s,jg s > is equivalent to p = q by applying (.. 5 The quaternion linear canonical transform The linear canonical transform (LCT, as a generalization of the classical Fourier transform, has more degrees of freedom than the FT and the FFT, but with similar computation cost as the conventional FT [3]. In [4], the authors generalized the classical LCT to the quaternionic algebra and deمحned QLCTs. In this section, we will only focus on two-sided (sandwich QLCT and right-sided QLCT. 5
16 5. The SQLCT in L (,H and its extension to L (,H ( ai b i Definition 5. LetA i = be areal matrix parametersuch thatdet(a i = c i d i for i =,. The two-sided (sandwich QLCT of f L (,H is defined by K A i (x,ω f(x,x K j A (x,ω dx dx, b,b 0, d d (L s f(ω,ω := eic ω f(d ω,x K j A (x,ω dx, b = 0,b 0, d Ki A (x,ω f(x,d ω e jc d ω dx, b 0,b = 0, d d e ic d ω f(d ω,d ω e jc d ω, b = b = 0. where and K i A (x,ω := K j A (x,ω := iπb e i( a b x b x ω + d b w, for b 0 (5. jπb e j( a b x b x ω + d b w, for b 0. (5. We only consider the case of b b 0. If b i 0 for i =,, then ( ( ai b i di b = i. (5.3 c i d i c i a i Lemma 5. Let f L (,H and h(x,x = e i a x b f(x,x e j a x b, then (L s f(ω,ω = e i d w ω b (F s h(, ω e j d w b. (5.4 ib b b jb Theorem 5.3 (Inversion of SQLCT If f,l s f L (,H and g(x,x = K i A then f(x,x = g(x,x for almost every (x,x. Proof. Applying Lemma 5. and Theorem 4.4 we have f(x,x = e i a x b h(x,x e j a x b ( = e i a b x = π = K i A = g(x,x. e i (ω,x (L s f(ω,ω K j (ω A,x dω dω, (5.5 ω b x (F s h( ω e jω b x dω dω e j a b x, ω π b b b b e i( d w b +ω x b a x b (L s f(ω,ω e j( d ib jb (ω,x (L s f(ω,ω K j (ω A,x dω dω w b +ω x b a b x dω dω 6
17 The second equality holds for almost every (x,x and the last equality is a consequence of (5., (5. and (5.3. So we can deمحne the inverse SQLCT L s by right-side of (5.5. Furthermore, L s (L s can be extended to L (,H. Theorem 5.4 (SQLCT on L (,H Letf,g L (,H, h(x,x := e i a x b f(x,x e j a x b, define (Φ s f(ω,ω := e i d w b ω (Ψ s h(, ω e j d w b. ib b b jb Then the following assertions hold.. Φ s is a bijection on L (,H. Moreover, Φ s and Φ s coincide with L s and L s on L (,H L (,H respectively.. Parseval sidentity Φ s f = f holds. Furthermore, Sc(< f,g > = Sc(< Φ s f,φ s g > and Sc(i < f,g > = Sc(i < Φ s f,φ s g >. 5. The QLCT in L (,H and its extension to L (,H Now we turn our study to right-sided QLCT. ( ai b i Definition 5.5 LetA i = be areal matrix parametersuch thatdet(a i = c i d i for i =,. The right-sided QLCT of f L (,H is defined by f(x,x KA i (x,ω K j A (x,ω dx dx, b,b 0, d (L r f(ω,ω := f(d ω,x e ic d ω K j A (x,ω dx, b = 0,b 0, d f(x,d ω KA i (x,ω e jc d ω dx, b 0,b = 0, d d f(d ω,d ω e ic d ω e j c d ω, b = b = 0. We want to build relationship between QLCT and transforms deمحned in former sections so that previous results could be employed in this section. However, due to the quadratic term of kernel (5. and (5., itلاs an extravagant hope to have a good relation like (4. for L r and general H-valued functions f. The two-dimensional QLCT can be considered as a cascade of two one-dimensional transforms. For any xedمح unit pure imaginary quaternion µ, let f(t L (,H, deمحne transform F µ of f as (F µ f(ξ = f(te µtξ dt. π By making a similar but easier proof to Theorem 3.3 and 3.0, we obtain the following Lemma. 7
18 Lemma 5.6 f,f µ f L (,H, then holds for almost every t. f(t = π (F µ f(ξe µtξ dξ Furthermore, F µ can be defined for every f L (,H and F µ f = f. To investigate right-sided QLCT, we will only be concerned with the case of b b 0 in following. Theorem 5.7 (Inversion of QLCT If f,l r f L (,H and f(x,x = (L r f(ω,ω K j (ω A,x K i (ω A,x dω dω, (5.6 holds for almost every (x,x. Proof. Let g x (x = g(x,x = f(x,x e i a x b, m ω (x = f(x,x KA i (x,ω dx and n ω (x = m ω (x e i a x b. Then ( ω (L r f(ω,ω = (F j n ω e j d ω b. (5.7 jb Since f,l r f L (,H. It is easy to verify that m ω,n ω,(l r f(ω, L (,H for almost every ω by Fubiniلاs theorem. Then by invoking Lemma 5.6, n ω (x = (L r f(ω,ω e j d ω b jb e jω x b b dω holds for almost every x. Note that Thus m ω (x = (F i g x ( ω b b e j d b ω ib. (5.8 f(x,x = g x (x e i a x b ( = n ω (x e i a x b e j d ω b ib e iω x b b dω = (L r f(ω,ω K j A (ω,x K i (ω A,x dω dω holds for almost every (x,x by invoking Lemma 5.6 again. e i a b x For every f L p (,H(p =,, we can decompose it into two parts: f = f i +jf i (5.9 where f i,f i are C i -valued. If we deمحne Φ r f := Φ s f i + jφ s f i, then Φ r is well-deمحned for every f L (,H and coincides with L r on L (,H L (,H. 8
19 Theorem 5.8 For every f,g L (,H, < f,g >=< Φ r f,φ r g > holds. In particular, Φ r f = f. Proof. If f L (,H L (,H then L r f = L s f i + jl s f i. Since L s f i,l s f i L (,H, thus L r f L (,H. Applying Lemma 5.6 and noticing that (5.7 and (5.8, we have and Therefore (L r f(ω,ω dω = n ω (x dω = = = = = ( (F ω jn ω b b dω ( (F ω jn ω b b dω (F j n ω (ω dω = n ω (x dx m ω (x dω ( (F ω ig x b b dω (F i g x (ω dω = g(x,x dx. (L r f(ω,ω dω dω = n ω (x dx dω = n ω (x dω dx = g(x,x dx dx, namely, L r f = f. Since Φ r is the extension of L r on L (,H, then Φ r f = f. Since Φ r is left H-linear, applying Parsevalلاs identity to f+ig = Φ r f+iφ r g, f +jg = Φ rf +jφ r g, f +kg = Φ rf +kφ r g respectively, then < f,g >=< Φ r f,φ r g > holds for every f,g L (,H. In Theorem 4.9, we conclude that if both f and g are C i -valued, then < f,g >=< Ψ s f,ψ s g >. But the analogous result for SQLCT is not reached yet in Theorem 5.4. In fact, if f is C i -valued, we have Φ r f = Φ s f. Then by Theorem 5.8, we obtain the following corollary which is a supplement for Theorem 5.4. Corollary 5.9 Suppose that f,g L (,H and both of them are C i -valued, then < f,g >=< Φ s f,φ s g >. 9
20 6 Discussions and Conclusions Due to the non-commutativity of multiplication of quaternions, there are at least eight types of QFTs and we only consider two typical types of them. How about the rest of QFTs?. The left-sided QFT (LQFT e iω x f( follows a similar pattern to QFT, with kernel moved to left side. As left-sided QFT is right H-linear, we only need to revise the deمحnition of inner product in L (,H to be < f,g > L (,H= f(x,x g(x,x dx dx. Then the results of QFT still hold for LQFT case.. Ifi,jaresubstituted intoµ,µ respectively, whereµ,µ beanytwoperpendicular unit pure imaginary quaternion, all of above results still hold. 3. The types of e µω x e µω x f(,f( e µω x e µω x and e µω x f( e µω x (single axis types obviously easier than the types in present paper (factored types. The single axis types of QFTs have similar properties to factor types of QFTs. Moreover, the proofs will be simpler. 4. The inversion theorem and Plancherel theorem of transforms e µ ω x µ ω x f( and f( e µ ω x µ ω x (dual axis types have not been worked out yet in this paper. Each type of QLCTs corresponds to a speciمحc type of QFTs. So above four facts about QFTs are suited for QLCTs as well. In this paper, we investigate the behaviors of the QFTs (except for dual axis types on the space L (,H. They are reversible under the suitable condition (the transformed function still integrable. The QFTs on L (,H have symmetric property, for example, QFT is unitary on L (,H. We are largely exploiting the relations between different types of QFTs. These relations are helpful to study the properties of QFTs. 7 Acknowledgements The authors acknowledge nancialمح support from the National Natural Science Funds for Young Scholars (No , 5005 and University of Macau No. MYG FST, MYG099(Y-L-FST3-KKI and the Macao Science and Technology Development Fund FDCT/094/0A, FDCT/099/0/A3. 0
21 eferences [] Mawardi Bahri, Eckhard MS Hitzer, Akihisa Hayashi, and yuichi Ashino. An uncer- تح 56(9:398 Appl., tainty principle for quaternion Fourier transform. Comput. Math. 40, 008. [] Patrick Bas, Nicolas Le Bihan, and Jean-Marc Chassery. Color image watermarking using quaternion fourier transform. In Acoustics, Speech, and Signal Processing, 003. Proceedings.(ICASSP IEEE International Conference on, volume 3, pages. 5 تحIII IEEE, 003. [3] Eduardo Bayro-Corrochano, Noel Trujillo, and Michel Naranjo. Quaternion fourier descriptors for the preprocessing and recognition of spoken words using images of spatiotemporal representations. Journal of Mathematical Imaging and Vision,.007, 90 تح 8(:79 [4] Nicolas Le Bihan and Stephen J Sangwine. Quaternion principal component analysis of color images. In Image Processing, 003. ICIP 003. Proceedings. 003 International Conference on, volume, pages. 809 تحI IEEE, 003. [5] Fred Brackx, ichard Delanghe, and Franciscus Sommen. Clifford Analysis, volume 76. Pitman Books Limited, 98. [6] Thomas Buغlow. Hypercomplex Spectral Signal epresentations for The Processing and Analysis of Images. Universitغat Kiel. Institut fuغr Informatik und Praktische Mathematik, 999. [7] Beijing Chen, Gouenou Coatrieux, Gang Chen, Xingming Sun, Jean Louis Coatrieux, and Huazhong Shu. Full 4-d quaternion discrete fourier transform based 04., 9 تح 8:06 Processing, watermarking for color images. Digital Signal [8] Todd Ell, Stephen J Sangwine, et al. Hypercomplex Fourier transforms of color 007., 35 تح 6(: Process., images. IEEE Trans. Image [9] Todd Anthony Ell. Quaternion-Fourier transforms for analysis of two-dimensional linear time-invariant partial differential systems. In Decision and Control, 993., Proceedings of the 3nd IEEE Conference on, pages. 84 تح 830 IEEE, 993. [0] Todd Anthony Ell. Quaternion Fourier transform: e-tooling image and signal processing analysis. In Quaternion and Clifford Fourier Transforms and Wavelets, pages. 4 تح 3 Springer, 03. [] S Georgiev, J Morais, KI Kou, and W. igظغspro Bochnerتحminlos theorem and quaternion fourier transform. In Quaternion and Clifford Fourier Transforms and Wavelets, pages. 0 تح 05 Springer, 03.
22 [] Eckhard MS Hitzer. Quaternion Fourier transform on quaternion eldsمح and gener- 007., 57 تح 7(3:497 Alg., alizations. Adv. Appl. Clifford [3] Aykut Koc, Haldun M Ozaktas, Cagatay Candan, and M Alper Kutay. Digital computation of linear canonical transforms. Signal Processing, IEEE Transactions 008., 394 تح 56(6:383 on, [4] Kit Ian Kou, Jianyu Ou, and Joao Morais. Uncertainty principles associated with quaternionic linear canonical transforms. Mathematical Methods in the Applied Sciences, 06. [5] Walter udin. eal and Complex Analysis. Tata McGraw-Hill Education, 987. [6] Stephen John Sangwine. Fourier transforms of colour images using quaternion or 996., 980 تح 3(:979 Lett., hypercomplex, numbers. Electron. [7] Elias M Stein and Guido L Weiss. Introduction to Fourier Analysis on Euclidean Spaces, volume. Princeton university press, 97.
LOGARITHMIC UNCERTAINTY PRINCIPLE FOR QUATERNION LINEAR CANONICAL TRANSFORM
Proceedings of the 06 International Conference on Wavelet Analysis and Pattern Recognition Jeju South Korea 0-3 July LOGARITHMIC UNCERTAINTY PRINCIPLE FOR QUATERNION LINEAR CANONICAL TRANSFORM MAWARDI
More informationAn Uncertainty Principle for Quaternion Fourier Transform
An Uncertainty Principle for Quaternion Fourier Transform Mawardi Bahri a, Eckhard S. M. Hitzer a Akihisa Hayashi a Ryuichi Ashino b, a Department of Applied Physics, University of Fukui, Fukui 9-857,
More informationResearch Article A Simplified Proof of Uncertainty Principle for Quaternion Linear Canonical Transform
Abstract and Applied Analysis Volume 06, Article ID 5874930, pages http://doiorg/055/06/5874930 Research Article A Simplified Proof of Uncertainty Principle for Quaternion Linear Canonical Transform Mawardi
More informationProduct Theorem for Quaternion Fourier Transform
Int. Journal of Math. Analysis, Vol. 8, 204, no. 2, 8-87 HIKAI Ltd, www.m-hikari.com http://dx.doi.org/0.2988/ijma.204.3290 Product Theorem for Quaternion ourier Transform Mawardi Bahri Department of Mathematics,
More informationDemystification of the Geometric Fourier Transforms
Demystification of the Geometric Fourier Transforms Roxana Buack, Gerik Scheuermann and Eckhard Hitzer Universität Leipzig, Institut für Informatik, Abteilung für Bild- und Signalverarbeitung, Augustuplatz
More informationShort-time Fourier transform for quaternionic signals
Short-time Fourier transform for quaternionic signals Joint work with Y. Fu and U. Kähler P. Cerejeiras Departamento de Matemática Universidade de Aveiro pceres@ua.pt New Trends and Directions in Harmonic
More informationSome results on the lattice parameters of quaternionic Gabor frames
Some results on the lattice parameters of quaternionic Gabor frames S. Hartmann Abstract Gabor frames play a vital role not only modern harmonic analysis but also in several fields of applied mathematics,
More informationTighter Uncertainty Principles Based on Quaternion Fourier Transform
Adv. Appl. Clifford Algebras 6 (016, 479 497 c 015 Springer Basel 0188-7009/010479-19 published online July, 015 DOI 10.1007/s00006-015-0579-0 Advances in Applied Clifford Algebras Tighter Uncertainty
More informationOn uncertainty principle for quaternionic linear canonical transform
On uncertainty principle for quaternionic linear canonical transform Kit-Ian Kou Department of Mathematics, Faculty of Science and Technology, University of Macau, Macau. Jian-Yu Ou Department of Mathematics,
More informationMotivation towards Clifford Analysis: The classical Cauchy s Integral Theorem from a Clifford perspective
Motivation towards Clifford Analysis: The classical Cauchy s Integral Theorem from a Clifford perspective Lashi Bandara November 26, 29 Abstract Clifford Algebras generalise complex variables algebraically
More informationThe orthogonal planes split of quaternions and its relation to quaternion geometry of rotations 1
The orthogonal planes split of quaternions and its relation to quaternion geometry of rotations 1 Eckhard Hitzer Osawa 3-10-, Mitaka 181-8585, International Christian University, Japan E-mail: hitzer@icu.ac.jp
More informationOn Riesz-Fischer sequences and lower frame bounds
On Riesz-Fischer sequences and lower frame bounds P. Casazza, O. Christensen, S. Li, A. Lindner Abstract We investigate the consequences of the lower frame condition and the lower Riesz basis condition
More informationArchiv der Mathematik Holomorphic approximation of L2-functions on the unit sphere in R^3
Manuscript Number: Archiv der Mathematik Holomorphic approximation of L-functions on the unit sphere in R^ --Manuscript Draft-- Full Title: Article Type: Corresponding Author: Holomorphic approximation
More informationbe the set of complex valued 2π-periodic functions f on R such that
. Fourier series. Definition.. Given a real number P, we say a complex valued function f on R is P -periodic if f(x + P ) f(x) for all x R. We let be the set of complex valued -periodic functions f on
More informationInternational Journal of Wavelets, Multiresolution and Information Processing c World Scientific Publishing Company
October 6, 014 1:7 WSPC/WS-IJWMIP QTF-ijwmip revf International Journal of Wavelets, Multiresolution and Information Processing c World Scientific Publishing Company CONTINUOUS QUATENION FOUIE AND WAVELET
More informationInfinite-dimensional Vector Spaces and Sequences
2 Infinite-dimensional Vector Spaces and Sequences After the introduction to frames in finite-dimensional vector spaces in Chapter 1, the rest of the book will deal with expansions in infinitedimensional
More informationConnecting spatial and frequency domains for the quaternion Fourier transform
Connecting spatial and frequency domains for the quaternion Fourier transform Ghent University (joint work with N. De Schepper, T. Ell, K. Rubrecht and S. Sangwine) MOIMA, Hannover, June, 2016 Direct formulas
More informationQUATERNIONS AND ROTATIONS
QUATERNIONS AND ROTATIONS SVANTE JANSON 1. Introduction The purpose of this note is to show some well-known relations between quaternions and the Lie groups SO(3) and SO(4) (rotations in R 3 and R 4 )
More informationCHAPTER VIII HILBERT SPACES
CHAPTER VIII HILBERT SPACES DEFINITION Let X and Y be two complex vector spaces. A map T : X Y is called a conjugate-linear transformation if it is a reallinear transformation from X into Y, and if T (λx)
More informationTwo-Dimensional Clifford Windowed Fourier Transform
Two-Dimensional Clifford Windowed Fourier Transform Mawardi Bahri, Eckhard M. S. Hitzer and Sriwulan Adji Abstract Recently several generalizations to higher dimension of the classical Fourier transform
More informationFinite-dimensional spaces. C n is the space of n-tuples x = (x 1,..., x n ) of complex numbers. It is a Hilbert space with the inner product
Chapter 4 Hilbert Spaces 4.1 Inner Product Spaces Inner Product Space. A complex vector space E is called an inner product space (or a pre-hilbert space, or a unitary space) if there is a mapping (, )
More informationElements of Positive Definite Kernel and Reproducing Kernel Hilbert Space
Elements of Positive Definite Kernel and Reproducing Kernel Hilbert Space Statistical Inference with Reproducing Kernel Hilbert Space Kenji Fukumizu Institute of Statistical Mathematics, ROIS Department
More informationKernel Method: Data Analysis with Positive Definite Kernels
Kernel Method: Data Analysis with Positive Definite Kernels 2. Positive Definite Kernel and Reproducing Kernel Hilbert Space Kenji Fukumizu The Institute of Statistical Mathematics. Graduate University
More informationEECS 598: Statistical Learning Theory, Winter 2014 Topic 11. Kernels
EECS 598: Statistical Learning Theory, Winter 2014 Topic 11 Kernels Lecturer: Clayton Scott Scribe: Jun Guo, Soumik Chatterjee Disclaimer: These notes have not been subjected to the usual scrutiny reserved
More informationOle Christensen 3. October 20, Abstract. We point out some connections between the existing theories for
Frames and pseudo-inverses. Ole Christensen 3 October 20, 1994 Abstract We point out some connections between the existing theories for frames and pseudo-inverses. In particular, using the pseudo-inverse
More informationChapter 8 Integral Operators
Chapter 8 Integral Operators In our development of metrics, norms, inner products, and operator theory in Chapters 1 7 we only tangentially considered topics that involved the use of Lebesgue measure,
More informationThe Fueter Theorem and Dirac symmetries
The Fueter Theorem and Dirac symmetries David Eelbode Departement of Mathematics and Computer Science University of Antwerp (partially joint work with V. Souček and P. Van Lancker) General overview of
More informationMatrices over Hyperbolic Split Quaternions
Filomat 30:4 (2016), 913 920 DOI 102298/FIL1604913E Published by Faculty of Sciences and Mathematics, University of Niš, Serbia Available at: http://wwwpmfniacrs/filomat Matrices over Hyperbolic Split
More informationMORE NOTES FOR MATH 823, FALL 2007
MORE NOTES FOR MATH 83, FALL 007 Prop 1.1 Prop 1. Lemma 1.3 1. The Siegel upper half space 1.1. The Siegel upper half space and its Bergman kernel. The Siegel upper half space is the domain { U n+1 z C
More informationC -Algebra B H (I) Consisting of Bessel Sequences in a Hilbert Space
Journal of Mathematical Research with Applications Mar., 2015, Vol. 35, No. 2, pp. 191 199 DOI:10.3770/j.issn:2095-2651.2015.02.009 Http://jmre.dlut.edu.cn C -Algebra B H (I) Consisting of Bessel Sequences
More informationThe orthogonal planes split of quaternions and its relation to quaternion geometry of rotations
Home Search Collections Journals About Contact us My IOPscience The orthogonal planes split of quaternions and its relation to quaternion geometry of rotations This content has been downloaded from IOPscience.
More informationClosed-Form Solution Of Absolute Orientation Using Unit Quaternions
Closed-Form Solution Of Absolute Orientation Using Unit Berthold K. P. Horn Department of Computer and Information Sciences November 11, 2004 Outline 1 Introduction 2 3 The Problem Given: two sets of corresponding
More informationON THE BEHAVIOR OF THE SOLUTION OF THE WAVE EQUATION. 1. Introduction. = u. x 2 j
ON THE BEHAVIO OF THE SOLUTION OF THE WAVE EQUATION HENDA GUNAWAN AND WONO SETYA BUDHI Abstract. We shall here study some properties of the Laplace operator through its imaginary powers, and apply the
More informationA Concise Course on Stochastic Partial Differential Equations
A Concise Course on Stochastic Partial Differential Equations Michael Röckner Reference: C. Prevot, M. Röckner: Springer LN in Math. 1905, Berlin (2007) And see the references therein for the original
More informationRené Bartsch and Harry Poppe (Received 4 July, 2015)
NEW ZEALAND JOURNAL OF MATHEMATICS Volume 46 2016, 1-8 AN ABSTRACT ALGEBRAIC-TOPOLOGICAL APPROACH TO THE NOTIONS OF A FIRST AND A SECOND DUAL SPACE III René Bartsch and Harry Poppe Received 4 July, 2015
More informationHeisenberg's inequality for Fourier transform
Heisenberg's inequality for Fourier transform Riccardo Pascuzzo Abstract In this paper, we prove the Heisenberg's inequality using the Fourier transform. Then we show that the equality holds for the Gaussian
More informationON MATRIX VALUED SQUARE INTEGRABLE POSITIVE DEFINITE FUNCTIONS
1 2 3 ON MATRIX VALUED SQUARE INTERABLE POSITIVE DEFINITE FUNCTIONS HONYU HE Abstract. In this paper, we study matrix valued positive definite functions on a unimodular group. We generalize two important
More informationLie Algebra of Unit Tangent Bundle in Minkowski 3-Space
INTERNATIONAL ELECTRONIC JOURNAL OF GEOMETRY VOLUME 12 NO. 1 PAGE 1 (2019) Lie Algebra of Unit Tangent Bundle in Minkowski 3-Space Murat Bekar (Communicated by Levent Kula ) ABSTRACT In this paper, a one-to-one
More informationRepresentation of objects. Representation of objects. Transformations. Hierarchy of spaces. Dimensionality reduction
Representation of objects Representation of objects For searching/mining, first need to represent the objects Images/videos: MPEG features Graphs: Matrix Text document: tf/idf vector, bag of words Kernels
More informationKaczmarz algorithm in Hilbert space
STUDIA MATHEMATICA 169 (2) (2005) Kaczmarz algorithm in Hilbert space by Rainis Haller (Tartu) and Ryszard Szwarc (Wrocław) Abstract The aim of the Kaczmarz algorithm is to reconstruct an element in a
More information4 Hilbert spaces. The proof of the Hilbert basis theorem is not mathematics, it is theology. Camille Jordan
The proof of the Hilbert basis theorem is not mathematics, it is theology. Camille Jordan Wir müssen wissen, wir werden wissen. David Hilbert We now continue to study a special class of Banach spaces,
More informationSPECTRAL THEOREM FOR COMPACT SELF-ADJOINT OPERATORS
SPECTRAL THEOREM FOR COMPACT SELF-ADJOINT OPERATORS G. RAMESH Contents Introduction 1 1. Bounded Operators 1 1.3. Examples 3 2. Compact Operators 5 2.1. Properties 6 3. The Spectral Theorem 9 3.3. Self-adjoint
More informationRELATIONSHIP BETWEEN QUATERNION LINEAR CANONICAL AND QUATERNION FOURIER TRANSFORMS
Proceedings of the 04 Internationa Conference on Waveet Anaysis and Pattern ecognition, Lanzhou, 3-6 Juy, 04 ELATIONSHIP BETWEEN QUATENION LINEA CANONICAL AND QUATENION FOUIE TANSFOMS MAWADI BAHI, YUICHI
More information1. Fourier Transform (Continuous time) A finite energy signal is a signal f(t) for which. f(t) 2 dt < Scalar product: f(t)g(t)dt
1. Fourier Transform (Continuous time) 1.1. Signals with finite energy A finite energy signal is a signal f(t) for which Scalar product: f(t) 2 dt < f(t), g(t) = 1 2π f(t)g(t)dt The Hilbert space of all
More informationBochner s Theorem on the Fourier Transform on R
Bochner s heorem on the Fourier ransform on Yitao Lei October 203 Introduction ypically, the Fourier transformation sends suitable functions on to functions on. his can be defined on the space L ) + L
More informationOn complexified quantum mechanics and space-time
On complexified quantum mechanics and space-time Dorje C. Brody Mathematical Sciences Brunel University, Uxbridge UB8 3PH dorje.brody@brunel.ac.uk Quantum Physics with Non-Hermitian Operators Dresden:
More informationFourier Transform & Sobolev Spaces
Fourier Transform & Sobolev Spaces Michael Reiter, Arthur Schuster Summer Term 2008 Abstract We introduce the concept of weak derivative that allows us to define new interesting Hilbert spaces the Sobolev
More information4.7 The Levi-Civita connection and parallel transport
Classnotes: Geometry & Control of Dynamical Systems, M. Kawski. April 21, 2009 138 4.7 The Levi-Civita connection and parallel transport In the earlier investigation, characterizing the shortest curves
More informationChapter One. The Calderón-Zygmund Theory I: Ellipticity
Chapter One The Calderón-Zygmund Theory I: Ellipticity Our story begins with a classical situation: convolution with homogeneous, Calderón- Zygmund ( kernels on R n. Let S n 1 R n denote the unit sphere
More informationBoolean Inner-Product Spaces and Boolean Matrices
Boolean Inner-Product Spaces and Boolean Matrices Stan Gudder Department of Mathematics, University of Denver, Denver CO 80208 Frédéric Latrémolière Department of Mathematics, University of Denver, Denver
More informationFolland: Real Analysis, Chapter 8 Sébastien Picard
Folland: Real Analysis, Chapter 8 Sébastien Picard Problem 8.3 Let η(t) = e /t for t >, η(t) = for t. a. For k N and t >, η (k) (t) = P k (/t)e /t where P k is a polynomial of degree 2k. b. η (k) () exists
More informationFunctional Analysis. Martin Brokate. 1 Normed Spaces 2. 2 Hilbert Spaces The Principle of Uniform Boundedness 32
Functional Analysis Martin Brokate Contents 1 Normed Spaces 2 2 Hilbert Spaces 2 3 The Principle of Uniform Boundedness 32 4 Extension, Reflexivity, Separation 37 5 Compact subsets of C and L p 46 6 Weak
More informationReal Analysis, 2nd Edition, G.B.Folland Elements of Functional Analysis
Real Analysis, 2nd Edition, G.B.Folland Chapter 5 Elements of Functional Analysis Yung-Hsiang Huang 5.1 Normed Vector Spaces 1. Note for any x, y X and a, b K, x+y x + y and by ax b y x + b a x. 2. It
More information08a. Operators on Hilbert spaces. 1. Boundedness, continuity, operator norms
(February 24, 2017) 08a. Operators on Hilbert spaces Paul Garrett garrett@math.umn.edu http://www.math.umn.edu/ garrett/ [This document is http://www.math.umn.edu/ garrett/m/real/notes 2016-17/08a-ops
More information1 Fourier Transformation.
Fourier Transformation. Before stating the inversion theorem for the Fourier transformation on L 2 (R ) recall that this is the space of Lebesgue measurable functions whose absolute value is square integrable.
More informationTools from Lebesgue integration
Tools from Lebesgue integration E.P. van den Ban Fall 2005 Introduction In these notes we describe some of the basic tools from the theory of Lebesgue integration. Definitions and results will be given
More informationSobolev spaces. May 18
Sobolev spaces May 18 2015 1 Weak derivatives The purpose of these notes is to give a very basic introduction to Sobolev spaces. More extensive treatments can e.g. be found in the classical references
More informationFrame expansions in separable Banach spaces
Frame expansions in separable Banach spaces Pete Casazza Ole Christensen Diana T. Stoeva December 9, 2008 Abstract Banach frames are defined by straightforward generalization of (Hilbert space) frames.
More informationWavelets, wavelet networks and the conformal group
Wavelets, wavelet networks and the conformal group R. Vilela Mendes CMAF, University of Lisbon http://label2.ist.utl.pt/vilela/ April 2016 () April 2016 1 / 32 Contents Wavelets: Continuous and discrete
More informationFunctional Analysis Exercise Class
Functional Analysis Exercise Class Week: January 18 Deadline to hand in the homework: your exercise class on week January 5 9. Exercises with solutions (1) a) Show that for every unitary operators U, V,
More informationMATH MEASURE THEORY AND FOURIER ANALYSIS. Contents
MATH 3969 - MEASURE THEORY AND FOURIER ANALYSIS ANDREW TULLOCH Contents 1. Measure Theory 2 1.1. Properties of Measures 3 1.2. Constructing σ-algebras and measures 3 1.3. Properties of the Lebesgue measure
More informationMath 350 Fall 2011 Notes about inner product spaces. In this notes we state and prove some important properties of inner product spaces.
Math 350 Fall 2011 Notes about inner product spaces In this notes we state and prove some important properties of inner product spaces. First, recall the dot product on R n : if x, y R n, say x = (x 1,...,
More informationI teach myself... Hilbert spaces
I teach myself... Hilbert spaces by F.J.Sayas, for MATH 806 November 4, 2015 This document will be growing with the semester. Every in red is for you to justify. Even if we start with the basic definition
More informationMATH 205C: STATIONARY PHASE LEMMA
MATH 205C: STATIONARY PHASE LEMMA For ω, consider an integral of the form I(ω) = e iωf(x) u(x) dx, where u Cc (R n ) complex valued, with support in a compact set K, and f C (R n ) real valued. Thus, I(ω)
More informationRKHS, Mercer s theorem, Unbounded domains, Frames and Wavelets Class 22, 2004 Tomaso Poggio and Sayan Mukherjee
RKHS, Mercer s theorem, Unbounded domains, Frames and Wavelets 9.520 Class 22, 2004 Tomaso Poggio and Sayan Mukherjee About this class Goal To introduce an alternate perspective of RKHS via integral operators
More informationElementary linear algebra
Chapter 1 Elementary linear algebra 1.1 Vector spaces Vector spaces owe their importance to the fact that so many models arising in the solutions of specific problems turn out to be vector spaces. The
More informationDual Spaces. René van Hassel
Dual Spaces René van Hassel October 1, 2006 2 1 Spaces A little scheme of the relation between spaces in the Functional Analysis. FA spaces Vector space Topological Space Topological Metric Space Vector
More informationA BRIEF INTRODUCTION TO HILBERT SPACE FRAME THEORY AND ITS APPLICATIONS AMS SHORT COURSE: JOINT MATHEMATICS MEETINGS SAN ANTONIO, 2015 PETER G. CASAZZA Abstract. This is a short introduction to Hilbert
More informationInequalities of Babuška-Aziz and Friedrichs-Velte for differential forms
Inequalities of Babuška-Aziz and Friedrichs-Velte for differential forms Martin Costabel Abstract. For sufficiently smooth bounded plane domains, the equivalence between the inequalities of Babuška Aziz
More informationWhere is matrix multiplication locally open?
Linear Algebra and its Applications 517 (2017) 167 176 Contents lists available at ScienceDirect Linear Algebra and its Applications www.elsevier.com/locate/laa Where is matrix multiplication locally open?
More informationClifford Algebras and Spin Groups
Clifford Algebras and Spin Groups Math G4344, Spring 2012 We ll now turn from the general theory to examine a specific class class of groups: the orthogonal groups. Recall that O(n, R) is the group of
More informationFourier-like Transforms
L 2 (R) Solutions of Dilation Equations and Fourier-like Transforms David Malone December 6, 2000 Abstract We state a novel construction of the Fourier transform on L 2 (R) based on translation and dilation
More informationAn introduction to some aspects of functional analysis
An introduction to some aspects of functional analysis Stephen Semmes Rice University Abstract These informal notes deal with some very basic objects in functional analysis, including norms and seminorms
More informationLet R be the line parameterized by x. Let f be a complex function on R that is integrable. The Fourier transform ˆf = F f is. e ikx f(x) dx. (1.
Chapter 1 Fourier transforms 1.1 Introduction Let R be the line parameterized by x. Let f be a complex function on R that is integrable. The Fourier transform ˆf = F f is ˆf(k) = e ikx f(x) dx. (1.1) It
More informationThe Dual of a Hilbert Space
The Dual of a Hilbert Space Let V denote a real Hilbert space with inner product, u,v V. Let V denote the space of bounded linear functionals on V, equipped with the norm, F V sup F v : v V 1. Then for
More informationOn Unitary Relations between Kre n Spaces
RUDI WIETSMA On Unitary Relations between Kre n Spaces PROCEEDINGS OF THE UNIVERSITY OF VAASA WORKING PAPERS 2 MATHEMATICS 1 VAASA 2011 III Publisher Date of publication Vaasan yliopisto August 2011 Author(s)
More informationLINEAR ALGEBRA BOOT CAMP WEEK 4: THE SPECTRAL THEOREM
LINEAR ALGEBRA BOOT CAMP WEEK 4: THE SPECTRAL THEOREM Unless otherwise stated, all vector spaces in this worksheet are finite dimensional and the scalar field F is R or C. Definition 1. A linear operator
More informationComplex Numbers and Quaternions for Calc III
Complex Numbers and Quaternions for Calc III Taylor Dupuy September, 009 Contents 1 Introduction 1 Two Ways of Looking at Complex Numbers 1 3 Geometry of Complex Numbers 4 Quaternions 5 4.1 Connection
More informationHilbert Spaces. Contents
Hilbert Spaces Contents 1 Introducing Hilbert Spaces 1 1.1 Basic definitions........................... 1 1.2 Results about norms and inner products.............. 3 1.3 Banach and Hilbert spaces......................
More information3. (a) What is a simple function? What is an integrable function? How is f dµ defined? Define it first
Math 632/6321: Theory of Functions of a Real Variable Sample Preinary Exam Questions 1. Let (, M, µ) be a measure space. (a) Prove that if µ() < and if 1 p < q
More informationAN ELEMENTARY PROOF OF THE SPECTRAL RADIUS FORMULA FOR MATRICES
AN ELEMENTARY PROOF OF THE SPECTRAL RADIUS FORMULA FOR MATRICES JOEL A. TROPP Abstract. We present an elementary proof that the spectral radius of a matrix A may be obtained using the formula ρ(a) lim
More informationMeasure Theory on Topological Spaces. Course: Prof. Tony Dorlas 2010 Typset: Cathal Ormond
Measure Theory on Topological Spaces Course: Prof. Tony Dorlas 2010 Typset: Cathal Ormond May 22, 2011 Contents 1 Introduction 2 1.1 The Riemann Integral........................................ 2 1.2 Measurable..............................................
More informationON THE BOUNDEDNESS BEHAVIOR OF THE SPECTRAL FACTORIZATION IN THE WIENER ALGEBRA FOR FIR DATA
ON THE BOUNDEDNESS BEHAVIOR OF THE SPECTRAL FACTORIZATION IN THE WIENER ALGEBRA FOR FIR DATA Holger Boche and Volker Pohl Technische Universität Berlin, Heinrich Hertz Chair for Mobile Communications Werner-von-Siemens
More informationHilbert Spaces. Hilbert space is a vector space with some extra structure. We start with formal (axiomatic) definition of a vector space.
Hilbert Spaces Hilbert space is a vector space with some extra structure. We start with formal (axiomatic) definition of a vector space. Vector Space. Vector space, ν, over the field of complex numbers,
More informationChapter 3: Baire category and open mapping theorems
MA3421 2016 17 Chapter 3: Baire category and open mapping theorems A number of the major results rely on completeness via the Baire category theorem. 3.1 The Baire category theorem 3.1.1 Definition. A
More informationDeterminant lines and determinant line bundles
CHAPTER Determinant lines and determinant line bundles This appendix is an exposition of G. Segal s work sketched in [?] on determinant line bundles over the moduli spaces of Riemann surfaces with parametrized
More informationNULL CONDITION FOR SEMILINEAR WAVE EQUATION WITH VARIABLE COEFFICIENTS. Fabio Catalano
Serdica Math J 25 (999), 32-34 NULL CONDITION FOR SEMILINEAR WAVE EQUATION WITH VARIABLE COEFFICIENTS Fabio Catalano Communicated by V Petkov Abstract In this work we analyse the nonlinear Cauchy problem
More informationSPACES ENDOWED WITH A GRAPH AND APPLICATIONS. Mina Dinarvand. 1. Introduction
MATEMATIČKI VESNIK MATEMATIQKI VESNIK 69, 1 (2017), 23 38 March 2017 research paper originalni nauqni rad FIXED POINT RESULTS FOR (ϕ, ψ)-contractions IN METRIC SPACES ENDOWED WITH A GRAPH AND APPLICATIONS
More informationNewtonian Mechanics. Chapter Classical space-time
Chapter 1 Newtonian Mechanics In these notes classical mechanics will be viewed as a mathematical model for the description of physical systems consisting of a certain (generally finite) number of particles
More informationThe Cylindrical Fourier Transform
The Cylindrical Fourier Transform Fred Brackx, Nele De Schepper, and Frank Sommen Abstract In this paper we devise a so-called cylindrical Fourier transform within the Clifford analysis context. The idea
More informationAnalysis-3 lecture schemes
Analysis-3 lecture schemes (with Homeworks) 1 Csörgő István November, 2015 1 A jegyzet az ELTE Informatikai Kar 2015. évi Jegyzetpályázatának támogatásával készült Contents 1. Lesson 1 4 1.1. The Space
More informationPatryk Pagacz. Characterization of strong stability of power-bounded operators. Uniwersytet Jagielloński
Patryk Pagacz Uniwersytet Jagielloński Characterization of strong stability of power-bounded operators Praca semestralna nr 3 (semestr zimowy 2011/12) Opiekun pracy: Jaroslav Zemanek CHARACTERIZATION OF
More informationDS-GA 1002 Lecture notes 0 Fall Linear Algebra. These notes provide a review of basic concepts in linear algebra.
DS-GA 1002 Lecture notes 0 Fall 2016 Linear Algebra These notes provide a review of basic concepts in linear algebra. 1 Vector spaces You are no doubt familiar with vectors in R 2 or R 3, i.e. [ ] 1.1
More informationWhen is the Ring of 2x2 Matrices over a Ring Galois?
International Journal of Algebra, Vol. 7, 2013, no. 9, 439-444 HIKARI Ltd, www.m-hikari.com http://dx.doi.org/10.12988/ija.2013.3445 When is the Ring of 2x2 Matrices over a Ring Galois? Audrey Nelson Department
More informationHilbert spaces. 1. Cauchy-Schwarz-Bunyakowsky inequality
(October 29, 2016) Hilbert spaces Paul Garrett garrett@math.umn.edu http://www.math.umn.edu/ garrett/ [This document is http://www.math.umn.edu/ garrett/m/fun/notes 2016-17/03 hsp.pdf] Hilbert spaces are
More informationThe following definition is fundamental.
1. Some Basics from Linear Algebra With these notes, I will try and clarify certain topics that I only quickly mention in class. First and foremost, I will assume that you are familiar with many basic
More informationDerivatives of Harmonic Bergman and Bloch Functions on the Ball
Journal of Mathematical Analysis and Applications 26, 1 123 (21) doi:1.16/jmaa.2.7438, available online at http://www.idealibrary.com on Derivatives of Harmonic ergman and loch Functions on the all oo
More informationAnalysis Preliminary Exam Workshop: Hilbert Spaces
Analysis Preliminary Exam Workshop: Hilbert Spaces 1. Hilbert spaces A Hilbert space H is a complete real or complex inner product space. Consider complex Hilbert spaces for definiteness. If (, ) : H H
More informationOctober 25, 2013 INNER PRODUCT SPACES
October 25, 2013 INNER PRODUCT SPACES RODICA D. COSTIN Contents 1. Inner product 2 1.1. Inner product 2 1.2. Inner product spaces 4 2. Orthogonal bases 5 2.1. Existence of an orthogonal basis 7 2.2. Orthogonal
More informationarxiv: v1 [math.gr] 8 Nov 2008
SUBSPACES OF 7 7 SKEW-SYMMETRIC MATRICES RELATED TO THE GROUP G 2 arxiv:0811.1298v1 [math.gr] 8 Nov 2008 ROD GOW Abstract. Let K be a field of characteristic different from 2 and let C be an octonion algebra
More information