Transcript intuition
MRA
basic concepts
Jyun-Ming Chen
Spring 2001
Introduction
• MRA (multiresolution analysis)
– Construct a hierarchy
of approximations to
functions in various
subspaces of a linear
vector space
• First explained in
finite-dimensional
linear vector space,
VN
• There exist N linearly
independent and
orthogonal basis
vectors
a1 , a 2 ,, a N
• Any vector in VN can be
expressed as a unique
linear combination of
these basis vectors
x 1a1 2a 2 N a N
Simple Illustration in Number
Representation
Space Vn : the set of numbers represente d
using n decimal digits
R; 3.1415926...
V5 : represent number using 101 ,10 0 ,10 1 ,10 2 , 10 3
in V5 , 3.142
V4 : represent number using 101 ,10 0 ,10 1 ,10 2
Properties :
in V4 , 3.14
V j V j 1 V j 2 ... R, that is
V1 : represent number using 101
in V5 , 0
lim Vn R
n
V 0i
i
Nested Vector Spaces
• VN 1 : Subspace of a lower dimension by
taking only N-1 of the N basis vectors, say
a1 , a 2 ,, a N-1
• Continuing, …
VN 2 ,VN 3 ,,V1
• Hence,
V1 V2 VN
Approximate a Vector in
Subspaces
• Best approximation: minimize discrepancy
e N-1 x N 1
• Let N 1 be its orthogonal projection of x
in the subspace
a3
a1
e
a2
Orthogonal Projection and
Least Square Error
Error vect or Ax b must be
perpendicu lar to the column space
E Ax b ; E 2 Ax b Ax b
T
dE 2
Least square solution :
0
dx
2 AT Ax b 0
Ay : Any vector in the column space
Ay T Ax b y T AT Ax b 0
For Orthonormal Basis (N=3)
a3
a3
x
a1
e2
1
a2
a1
e1
2 x, a1 a1 x, a 2 a 2
1 1 , a1 a1
e2 x 2
e1 2 1
x e2 e1 1
a2
Interpretations
• Approximating
vectors:
– Sequence of
orthogonal projection
vectors of x in the
subspaces
– Finest approximation
at VN-1
– Coarsest
approximation at V1
• Error (detail) vector in VN
– orthogonal to VN-1
• WN-1: the orthogonal
complement to VN-1 in VN
– dimensionality of 1
• Similarly, WN-2: the
orthogonal complement to
VN-2 in VN-1
Interpretations (cont)
• Every vector in VN can be
written in the form below (the
sum of one vector apiece
from the N subspaces WN-1,
WN-2, …, W1, V1)
x e N-1 e N-2 e1 1
• The vectors eN-1, eN-2, …, e1,
X1 form an orthogonal set
• VN is the direct sum of these
subspaces
VN WN 1 WN 2 W1 V1
V3
• Orthogonality of subspaces
Wk Vk
k 1, , N 1
Also
Wk W j , j 1,..., k 1
W2
V2
V1
W1
From Vector Space to
Function Space
Example of an MRA
l 1
f 0 (t ) f ( )d , l t l 1
• Let f(t) be a continuous,
l
1 2l 2
real-valued, finite
f 1 (t )
f ( )d , 2l t 2l 2
2
l
2
energy signal
1 l / 21/ 2
f ( )d , l / 2 t l / 2 1 / 2
• Approximate f(t) as f1 (t )
l
/
2
1/ 2
follows:
MRA Example (cont)
• V0: linear vector space, • Nested subspaces
formed by the set of
V1 V0 V1 V2
functions that are
piecewise constant
over unit interval
f (t ) f (t ) lim f k (t )
k
Approximating Function by
Orthogonal Projection
• Assume u is not a member of the space V spanned
by {fk}, a set of orthonormal basis
• We wish to find an approximation of u in V
u p u, f k f k
k
• Remarks:
– Approximation error u-up is orthogonal to space V
u u p , f k 0 k
– Mean square error of such an approximation is
minimum
Formal Definition of an MRA
An MRA consists of the nested linear vector space
such that V1 V0 V1 V2
• There exists a function f(t) (called scaling function)
such that f (t k ) : k integer is a basis for V0
• If f (t ) Vk then f (2t ) Vk 1 and vice versa
• lim V j L2 ( R) ; V j {0}
j
• Remarks:
– Does not require the set of f(t) and its integer translates
to be orthogonal (in general)
– No mention of wavelet