<< Chapter < Page Chapter >> Page >

We now look at how well f X can be approximated by n functions in the dictionary D .

We define the error of n -term approximation of f by the elements of the dictionary D as (1) σ n ( f ) X : = σ n ( f , D ) X : = inf s Σ n f - s X .

We also define the class of r -smooth signals in D as (2) A r : = A r ( D ) : = { f X , σ n ( f ) M n - r  for some  M } ,

with the corresponding norm f A r = sup n = 1 , 2 , n r σ n ( f ) X .

In general, the larger r is, the 'smoother' the function s A r ( D ) X . Note also that A r A r if r > r . Given f , let r ( f ) = sup { r : f A r } be a measure of the "smoothness" of f , i.e. a quantification of compressibility.

Let X = H , a Hilbert space

A Hilbert space is a complete inner product space with the norm induced by the innerproduct
such as X = L 2 ( R ) , and assume D = B - an orthonormal basis on X ; i.e. if B = { ϕ i } i , then ϕ i , ϕ j = δ i , j , where δ i , j is the Kronecker delta. This also means that each f X has an expansion f = j c j ( f ) ϕ j , where c j ( f ) = f , ϕ j . We also have f X 2 = j = 1 | c j ( f ) | 2 .

Recall the definition of p spaces: let ( a j ) R ; then ( a j ) p if ( a j ) p < with ( a j ) p = ( j | a j | p ) 1 / p for p < and ( a j ) p = sup j | a j | for p = . We also recall that for L p spaces on compact sets, L p L p if p > p . The opposite is true for p spaces: p p if p < p . Hence, the smaller the value of p is, the “smaller” p is.

Does there exist a sequence ( a j ) with ( a j ) 1 = j | a j | < but with ( a j ) p = ( j | a j | p ) 1 p = for all 0 < p < 1 ? Consider the sequence a n = 1 n ( log n ) 1 + δ . We see that ( a n ) 1 but ( a n ) p = for all 0 < p < 1 .

A sequence ( a n ) is in p if the sorted magnitudes of the a n decay faster than n - 1 p .

Define a n * as the element of the sequence ( a n ) with the n th largest magnitude, and denote ( a n * ) as the decreasing rearrangement of ( a n ) . It is easy to show that k ( a k * ) p n ( a n ) p for all k ; also, if ( a n ) p , then a k * ( a n ) p k - 1 p .

A sequence ( a n ) is in weak p , denoted ( a n ) w p , if a k * M k - 1 p . We also define the quasinorm

A quasinorm is has the properties of a norm except thatthe triangle inequality is replaced by the condition x + y C 0 [ x + y ] for some absolute constant C 0 .
( a n ) w p as the smallest M > 0 such that a k * M k - 1 p for each k .

The sequence a n = 1 n is in weak 1 but not in 1 .

For p , p such that p > p , we have p w p p .

Let D = B be an orthonormal basis for the Hilbert space X = H . For f X with representation in B = [ ϕ 1 , ϕ 2 , ] as f = n c n ( f ) ϕ n , we have f A r ( B ) X if and only if the sequence ( c n ( f ) ) w τ , with 1 τ = r + 1 2 . Moreover, there exist C 0 , C 0 R such that C 0 ( c n ( f ) ) w τ f A r C 0 ( c n ( f ) ) w τ .

Let r = 1 2 . f A 1 2 if and only if ( c n ( f ) ) w τ , i.e. if c n * ( f ) M n - 1 = M n .

We prove the converse statement; the forward statement proof is left to the reader. We would like to show thatif ( c n ( f ) ) w τ , then f A r , with r = 1 τ - 1 2 . The best n -term approximation of f in B is of the form s = k Λ a k ϕ k , Λ n . Therefore, we have: (3) σ n ( f ) X = inf s Σ n f - s X = inf s Σ n k Λ ( c k ( f ) - a k ) ϕ k + k Λ c k ( f ) ϕ k X = inf s Λ k Λ ( c k ( f ) - a k ) 2 + k Λ ( c k ( f ) ) 2 = k = n + 1 | c k * ( f ) | 2 M 2 k = n + 1 k - 2 τ M 2 k = n + 1 k - 2 r - 1  (since  ( c n ( f ) ) w p ) ,

where M : = ( c n ( f ) w p .

We prove the converse statement; the forward statement proof is left to the reader. We would like to show thatif ( c n ( f ) ) w τ , then f A r , with r = 1 τ - 1 2 . The best n -term approximation of f in B is of the form s = k Λ a k ϕ k , Λ n . Therefore, we have: (3) σ n ( f ) X = inf s Σ n f - s X = inf s Σ n k Λ ( c k ( f ) - a k ) ϕ k + k Λ c k ( f ) ϕ k X = inf s Λ k Λ ( c k ( f ) - a k ) 2 + k Λ ( c k ( f ) ) 2 = k = n + 1 | c k * ( f ) | 2 M 2 k = n + 1 k - 2 τ M 2 k = n + 1 k - 2 r - 1  (since  ( c n ( f ) ) w p ) ,

where we define C = λ 0 r . Using this result in the earlier statement, we get (5) k = n + 1 | c k * ( f ) | 2 C M 2 n - 2 r M 2 z n - r ;

this implies by definition that ( c k ( f ) ) A r .

Get Jobilize Job Search Mobile App in your pocket Now!

Get it on Google Play Download on the App Store Now




Source:  OpenStax, Compressive sensing. OpenStax CNX. Sep 21, 2007 Download for free at http://cnx.org/content/col10458/1.1
Google Play and the Google Play logo are trademarks of Google Inc.

Notification Switch

Would you like to follow the 'Compressive sensing' conversation and receive update notifications?

Ask