sparse image representation via combined transforms - Convex ...
sparse image representation via combined transforms - Convex ...
sparse image representation via combined transforms - Convex ...
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
30 CHAPTER 3. IMAGE TRANSFORMS AND IMAGE FEATURES<br />
Also let DFT N denote the N-point DFT matrix<br />
DFT N = {e −i 2π N kl } k=0,1,2,... ,N−1; ,<br />
l=0,1,2,... ,N−1<br />
where k is the row index and l is the column index. We have<br />
⃗̂X = DFT N · ⃗X.<br />
Apparently DFT N is an N by N square symmetric matrix. Both X ⃗ and ⃗̂X are N-<br />
dimensional vectors.<br />
It is well known that the matrix DFT N is orthogonal, which is equivalent to saying<br />
that the inverse of DFT N is its complex conjugate transpose. Thus, DFT is an orthogonal<br />
transform and consequently it is an isometric transform. This result is generally attributed<br />
to Parseval.<br />
Since the concise statements and proofs can be found in many textbooks, we skip the<br />
proof.<br />
Why Fourier Transform?<br />
The preponderant reason that Fourier analysis is so powerful in analyzing linear time invariant<br />
system and cyclic-stationary time series is the fact that Fourier series are the eigenvectors<br />
of matrices associated with these <strong>transforms</strong>. These <strong>transforms</strong> are ubiquitous in DSP and<br />
time series analysis.<br />
Before we state the key result, let us first introduce some mathematical notation:<br />
Toeplitz matrix, Hankel matrix and circulant matrix.<br />
Suppose A N×N is an N by N real-valued matrix: A = {a ij } 1≤i≤N,1≤j≤N with all a ij ∈ R.<br />
The matrix A is a Toeplitz matrix when the elements on the diagonal, and the rows that<br />
are parallel to the diagonal, are the same, or mathematically a ij = t i−j [74, page 193]. The<br />
following is a Toeplitz matrix:<br />
⎛<br />
T =<br />
⎜<br />
⎝<br />
⎞<br />
t 0 t −1 ... t −(n−2) t −(n−1)<br />
t 1 t 0 ... t −(n−3) t −(n−2)<br />
. . . .. . .<br />
.<br />
t n−2 t n−3 ... t 0 t −1<br />
⎟<br />
⎠<br />
t n−1 t n−2 ... t 1 t 0<br />
N×N