sparse image representation via combined transforms - Convex ...

sparse image representation via combined transforms - Convex ... sparse image representation via combined transforms - Convex ...

convexoptimization.com
from convexoptimization.com More from this publisher
10.03.2015 Views

78 CHAPTER 3. IMAGE TRANSFORMS AND IMAGE FEATURES Equations (3.38) (3.39) and (3.40) together are equivalent to the matrix multiplication ( H(z) H(−z) )( H(z −1 ) H(−z −1 ) ) ( 2 0 ) G(z) G(−z) G(z −1 ) G(−z −1 ) = 0 2 . Taking the determinant of both sides, we have [H(z)G(−z) − G(z)H(−z)][H(z −1 )G(−z −1 ) − G(z −1 )H(−z −1 )] = 4. (3.41) If both sequence {h(n) :n ∈ Z} and sequence {g(n) :n ∈ Z} have finite length (thinking of the impulse responses of FIR filters), then the polynomial H(z)G(−z) − G(z)H(−z) must have only one nonzero term. Since if the polynomial H(z)G(−z) − G(z)H(−z) has more than two terms and simultaneously has finite length, (3.41) can never be equal to a constant. On the other hand, we have [H(z −1 )+H(−z −1 )][H(z)G(−z) − G(z)H(−z)] = H(z −1 )H(z)G(−z)+H(−z −1 )H(z)G(−z) −H(z −1 )G(z)H(−z) − H(−z −1 )G(z)H(−z) (3.40) = H(z −1 )H(z)G(−z) − H(z −1 )H(z)G(z) +H(−z −1 )G(−z)H(−z) − H(−z −1 )G(z)H(−z) (3.38) = 2[G(−z) − G(z)]. We know that H(z)G(−z) − G(z)H(−z) only has one nonzero term. The polynomial H(z −1 )+H(−z −1 ) can only have terms with even exponents and polynomial G(−z) − G(z) can only have terms with odd exponents, so from the previous equation, there must exist an integer k, such that H(z)G(−z) − G(z)H(−z) =2z 2k+1 , and [H(z −1 )+H(−z −1 )]z 2k+1 = G(−z) − G(z). (3.42)

3.7. PROOFS 79 Similarly, we have [H(z −1 ) − H(−z −1 )][H(z)G(−z) − G(z)H(−z)] = 2[G(−z)+G(z)]. Consequently, [H(z −1 ) − H(−z −1 )]z 2k+1 = G(−z)+G(z). (3.43) From (3.42) and (3.43), we have G(z) =−z 2k+1 H(−z −1 ). The above equation is equivalent to relation (3.30) in the Theorem.

3.7. PROOFS 79<br />

Similarly, we have<br />

[H(z −1 ) − H(−z −1 )][H(z)G(−z) − G(z)H(−z)] = 2[G(−z)+G(z)].<br />

Consequently,<br />

[H(z −1 ) − H(−z −1 )]z 2k+1 = G(−z)+G(z). (3.43)<br />

From (3.42) and (3.43), we have<br />

G(z) =−z 2k+1 H(−z −1 ).<br />

The above equation is equivalent to relation (3.30) in the Theorem.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!