Wavelet Coding and Related Issues

ENEE631 Spring09
Lecture 13 (3/11/2009)
Wavelet Coding and Related Issues

Spring 09 Instructor: Min Wu
Electrical and Computer Engineering Department,
University of Maryland, College Park
bb.eng.umd.edu (select ENEE631 S09)
minwu@eng.umd.edu
M. Wu: ENEE631 Digital Image Processing (Spring'09)
UMCP ENEE631 Slides (created by M.Wu 2004)
Overview and Logistics
Last Time:
JPEG Compression
Subband and Wavelet based compression
Subband decomposition; basic ideas of EZW encoding
Today
Continue on Subband and Wavelet based compression
Exploit the structures between coefficients for removing redundancy

More on wavelet transform, and consideration on filters
Assignment#3 is Due Friday noon March 13

JPEG part may be turned in after spring break
Project#1 Posted. Due Monday April 6

Lec13 Wavelet Coding, etc. [2]
Review: 2-D Subband/DWT

Decomposition
Separable transform by successively
applying 1-D DWT to rows and columns
From Usevitchs article in

IEEE Sig.Proc. Mag. 9/01
Review: Key Concepts in EZW
Exploit multi-resolution and self-similarity
Parent-children relation among coeff.

Each parent coeff at level k spatially correlates with
4 coeff at level (k-1) of same orientation
A coeff at lowest band correlates with 3 coeff.
Coding significance map via zero-tree

Encode insignificance map w/ zero-trees
Avoid large overhead by coding only significant

coefficients
Successive approximation quantization

Send most-significant-bits first and gradually
refine coefficient value
Embedded nature of coded bit-stream
get higher fidelity image by adding extra
refining bits
=> Quality Scalability with fine granularity
From Usevitch (IEEE

Sig.Proc. Mag. 9/01)
Recall: EZW Algorithm and Example

Initial threshold ~ 2 ^ floor(log2 xmax)

Put all coeff. in dominant list
Dominant Pass (zig-zag across bands)

Assign symbol to each coeff. and
entropy encode symbols
ps positive significance
ns negative significance
iz isolated zero
ztr zero-tree root
Significant coefficients
Move to subordinate list

Set them to zero in dominant list
Subordinate Pass
Fig. From Usevitch (IEEE Sig. Proc. Mag. 9/01)
Output one bit for subordinate list
According to position in up/low half of quantization interval
Repeat with half threshold, until bit budget achieved

E.g. first look for coeff. above 32; then focus on coeff. between 16 and 32
Divide quantization interval of [32,64) by half:

[32,48) => 40 and [48, 64) => 56
After 1st Pass
From Usevitch (IEEE Sig.Proc. Mag. 9/01)
After 2nd Pass
[16,24) => 20, [24, 32) => 28;

[32,40) => 36, [40,48) => 44;
[48, 56) => 52, [56, 64) => 60
2nd Subordinate Pass:
orig. 53 symbol 0 reconstruct 52
34
0
36
-22
-20
21
20
UMCP ENEE631 Slides (created by M.Wu 2001/2004)
EZW and Beyond
Can apply DWT to entire images or larger blocks than 8x8
Symbol sequence can be entropy encoded

e.g. arithmetic coding
Cons of EZW
Poor error resilience; Difficult for selective spatial decoding
SPIHT (Set Partitioning in Hierarchal Trees)

Further improvement over EZW to remove redundancy
Achieve equal/better performance than EZW w/o arithmetic coding
EBCOT (Embedded Block Coding with Optimal Truncation)

Used in JPEG 2000
Address the shortcomings of EZW (random access, error resilience, )
Embedded wavelet coding in each block + bit-allocations among blocks
A Close Look at Wavelet Transform

Haar Transform unitary
Orthonormal Wavelet Filters
Biorthogonal Wavelet Filters
Construction of Haar Functions
Unique decomposition of integer k (p, q)

k = 0, , N-1 with N = 2n, 0 <= p <= n-1
q = 0, 1 (for p=0); 1 <= q <= 2 p (for p>0)
e.g., k=0
(0,0)
k=1
(0,1)
k=2
(1,1)
k=3
(1,2)
1
for x [0,1]
N
q- 12
1 p/2
q-1
2
for p x p
2
2
N
q- 12
1 p/2
q
2
for p x p
2
2
N
h0 ( x) h0, 0 ( x)
hk ( x) h p ,q ( x)
for other x [0,1]
k = 2p + q 1
remainder
k=4
(2,1)
hk(x) = h p,q(x) for x [0,1]
power of 2
(0,0)
x
1
(0,1)
(1,1)
(1,2)
(2,1)
Haar Transform
Haar transform H
Sample hk(x) at {m/N}
1
1
for m = 0, , N-1
2
2 2 2
0
0
0
0
2 2 0
0
0 0
2 2
0 0
0
0
1
8
Real and orthogonal

Transition at each scale p is
localized according to q
0
2
1
1
0
0
0
2 2 2
0 0 0 0
0 0 0 0
2 2 0 0
2 2
Basis images of 2-D

(separable) Haar transform
Outer product of two basis vectors
Compare Basis Images of DCT and Haar
See also: Jains Fig.5.2 pp136

Time-Freq (or Space-Freq) Interpretations

Standard Basis (for data samples); Fourier Basis; Wavelet Basis
Recap on Haar Transform

Two major sub-operations

Scaling captures info. at different frequencies
Translation captures info. at different locations
Can be represented by filtering and downsampling
Relatively poor energy compaction

Equiv. filter response doesnt have good cutoff & stopband attenuation
x
1
Orthonormal Filters
Equiv. to projecting input signal to orthonormal basis
Energy preservation property

Convenient for quantizer design
MSE by transform domain quantizer is same as reconstruction MSE

in image domain
Shortcomings: coefficient expansion

Linear filtering with N-element input & M-element filter
(N+M-1)-element output (N+M)/2 after downsample

Length of output per stage grows ~ undesirable for compression
Solutions to coefficient expansion

Symmetrically extended input (circular convolution) &
Symmetric filter
Solutions to Coefficient Expansion

Circular convolution in place of linear convolution

Periodic extension of input signal
Problem: artifacts by large discontinuity at borders
Symmetric extension of input

Reduce border artifacts (note the signal length doubled with symmetry)
Problem: output at each stage may not be symmetric
From Usevitch (IEEE
Sig.Proc. Mag. 9/01)
Solutions to Coefficient Expansion (contd)
Symmetric extension + symmetric filters

No coefficient expansion and little artifacts
Symmetric filter (or asymmetric filter) => linear phase filters
(no phase distortion except by delays)
Problem
Only one set of linear phase filters for real FIR orthogonal wavelets
Haar filters: (1, 1) & (1,-1)
do not give good energy compaction
Ref: review ENEE630 discussions on FIR perfect reconstruction
Qudrature Mirror Filters (QMF) for 2-channel filter banks.
Biorthogonal Wavelets
Biorthogonal
Basis in forward and inverse transf.
are not the same but give overall
perfect reconstruction (PR)
recall EE630 PR filterbank
No strict orthogonality for transf.

filters so energy is not preserved
But could be close to orthogonal

filters performance
Advantage
Covers a much broader class of filters
including symmetric filters that eliminate coefficient expansion
Commonly used filters for compression

9/7 biorthogonal symmetric filter
Efficient implementation: Lifting approach
(ref: Sweldens tutorial)

Smoothness Conditions on Wavelet Filter

Ensure the low band coefficients obtained by recursive filtering
can provide a smooth approximation of the original signal
From M. Vetterlis wavelet/filter-bank paper

JPEG 2000 Image Compression Standard
From Christopoulos
(IEEE Trans. on CE
11/00)
JPEG 2000: A Wavelet-Based New Standard
Targets and features

Excellent low bit rate performance without sacrifice performance
at higher bit rate
Progressive decoding to allow from lossy to lossless
Region-of-interest (ROI) coding
Error resilience
For details
JPEG2000 Tutorial by Skrodras @ IEEE Sig. Proc Magazine 9/2001
David Taubman: High Performance Scalable Image Compression
with EBCOT, IEEE Trans. On Image Proc, vol.9(7), 7/2000.
Taubmans book on JPEG 2000 (on library reserve)
Links and tutorials @ http://www.jpeg.org/JPEG2000.htm
Examples
JPEG2K
vs.
JPEG
From Christopoulos
(IEEE Trans. on CE
11/00)
DCT vs. Wavelet: Which is Better?
3dB improvement?
Wavelet compression was claimed to have 3dB improvement over
DCT-based compression
Comparison is done on JPEG Baseline
Improvement not all due to transforms

Main contribution from better rate allocation, advanced entropy
coding, & smarter redundancy reduction via zero-tree
Difference due to DCT vs. Wavelet transform is less than 1dB
DCT coder can be improved to decrease the gap
[Ref] "A comparative study of DCT- and wavelet-based image coding",
Z. Xiong, K. Ramchandran, M. Orchard, Y-Q. Zhang, IEEE Trans. on Circuits and
Systems for Video Technology, v.9, no.5, 8/99, pp692-695.
Non-Dyadic Decomposition Wavelet Packets

Figures are from slides at Gonzalez/ Woods
DIP book 2/e website (Chapter 7)
Bit Allocation in Image Coding
Focus on MSE-based optimization;

Can further adjust based on HVS
Bit Allocation
How many bits to be allocated for each coefficient?

Related to each coeffs variance (and probability distribution)
More bits for high variance k2 to keep total MSE small
1
min D
N
1
subject to
N
E
N 1
k 0
N 1
r
k 0
v ( k ) vq ( k )
N 1
f (r )
k 0
R bits/coeff.
Here k2: variance of k-th coeff. v(k); rk is # bits allocated for v(k)
f(-): a func. relating bit rate to distortion based on a coeffs p.d.f.
Reverse Water-filling
Try to keep same amount of error in each frequency band
Rate/Distortion Allocation via Reverse Water-filling
Optimal solution for Gaussian distributed coefficients

Idea: try to keep same amount of error in each freq. band; no need to
spend bits to represent coeff. w/ smaller variance than water level
Results based on R-D function and via Lagrange-multiplier optimization
Note the convex shape of R-D function for Gaussian
=> slope becomes milder; require more bits to further reduce distortion
RD
Optimal
Not Optimal
D
2
Given total rate R, determine and then

D; or vice versa given D, determine R
Details on Reverse Water-filling Solution

RD
min Di ( Ri ) subj. to Ri R
i 1
Construct func. using

Lagrange multiplier
Necessary condition
Keep the same
marginal gain
J ( R1 ,...RN ) Di ( Ri ) R Ri
i 1
i 1
J dDi
dDi
0 for all i
and
Ri dRi
dRi
R R
i
Equal Slope Condition for bit allocation
1 i2
dRi
d
max ln
,0
dDi dDi
2 Di
dRi
1
1
for Ri 0
dDi
2 Di
Resulting in equal distortions on Gaussian r.vs

as interpreted in reverse water filling
Key Result in Rate Allocation: Equal R-D Slope

Rate allocation results:
=> Equal slope
(o.w. exist a better
allocation strategy)
Keep the slope in R-D curve the same (not just for Gaussian r.v)
Otherwise some bits can be applied to other r.v. for better return in
reducing the overall distortion
If all r.v. are Gaussian, the same slope in R-D curves for these r.v.
correspond to an identical amount of distortion.
R(D) = [ log 2 log D ] => dR/dD = 1 / (2D)
Use operational/measured rate-distortion information for general r.v.

Summary of Todays Lecture

Wavelet transform and coding

Haar basis/transform; considerations for wavelet filters; JPEG 2000
Optimal bit allocation: equal slope in R-D curves

Lead to reverse water filling for Gaussian coefficients
Next lecture
Optimal transform
Readings: see wavelet coding references in Project #1

Gonzalezs 3/e book Section 8.2.10; 7.1, 7.4-7.5; (7.2-7.3 on wavelet)
Transform coding articles in IEEE Sig. Proc. Magazine Sept. 2001
(special issue on transform coding)
To explore more: Tutorial on rate-distortion by Ortega-Ramchandran in IEEE
Sig. Proc. Magazine, Nov. 98, and Jains book Section 2.13;
Wavelets and the lifting scheme: A 5 minute tour by Wim Sweldens [Zeitschrift fr
Angewandte Mathematik und Mechanik, 76 (2), pp. 41-44, 1996]
Review: Correlation After a Linear Transform

Consider an Nx1 zero-mean random vector x

Covariance (or autocorrelation) matrix Rx = E[ x xH ]
give us ideas of the correlation between elements

Rx is a diagonal matrix for if all N r.v.s are uncorrelated
Apply a linear transform to x: y = A x
What is the autocorrelation matrix for y ?

Ry = E[ y yH ] = E[ (Ax) (Ax)H ] = E[ A x xH AH ]
= A E[ x xH ] AH = A Rx AH
Decorrelation: try to search for A that can produce a

decorrelated y (equiv. a diagonal correlation matrix Ry )
Optimal Transform
Recall: Why use transform in coding/compression?

Decorrelate the correlated data
Pack energy into a small number of coefficients
Interested in unitary/orthogonal or approximate orthogonal transforms
Energy preservation s.t. quantization effects can be better understood and

controlled
Unitary transforms weve dealt so far are data independent

Transform basis/filters are not depending on the signals we are processing
Can hard-wire them in encoder and decoder
What unitary transform gives the best energy compaction

and decorrelation?
Optimal in a statistical sense to allow the codec works well with many
images => Signal statistics would play an important role
Review: Correlation After a Linear Transform

Consider an Nx1 zero-mean random vector x

Covariance (or autocorrelation) matrix Rx = E[ x xH ]
give ideas of correlation between elements

Rx is a diagonal matrix for if all N r.v.s are uncorrelated
Apply a linear transform to x: y = A x
What is the correlation matrix for y ?

Ry = E[ y yH ] = E[ (Ax) (Ax)H ] = E[ A x xH AH ]
= A E[ x xH ] AH = A Rx AH
Decorrelation: try to search for A that can produce a

decorrelated y (equiv. a diagonal correlation matrix Ry )
K-L Transform (Principal Component Analysis)
Eigen decomposition of Rx:
Rx uk = k uk
Recall the properties of Rx
Hermitian (conjugate symmetric RH = R);

Nonnegative definite (real non-negative eigen values)
Karhunen-Loeve Transform (KLT)

y = UH x x = U y with U = [ u1, uN ]
KLT is a unitary transform with basis vectors in U being the
orthonormalized eigenvectors of Rx
Rx U = [ 1u1, N uN ] = U diag{1, 2, , N}
=> UH Rx U = diag{1, 2, , N} i.e. KLT performs decorrelation
Often order {ui} so that 1 2 N
Also known as Hotelling transform or Principle Component Analysis (PCA)

Properties of K-L Transform
Decorrelation
E[ y yH ]= E[ (UH x) (UH x)H ]= UH E[ x xH ] U = diag{1, 2, , N}
Note: Other matrices (unitary or nonunitary) may also decorrelate

the transformed sequence [Jains e.g.5.7 pp166]
Minimizing MSE under basis restriction

If only allow to keep m coefficients for any 1 m N, whats the
best way to minimize reconstruction error?
Keep the coefficients w.r.t. the eigenvectors of the first m largest
eigen values
Reference: Theorem 5.1 and Proof in Jains Book (pp166)
KLT Basis Restriction

Basis restriction
Keep only a subset of m transform coefficients and then perform
inverse transform (1 m N)
Basis restriction error: MSE between original & new sequences
Goal: to find the forward and backward transform matrices

to minimize the restriction error for each and every m
The minimum is achieved by KLT arranged according to the
decreasing order of the eigenvalues of R
K-L Transform for Images
Work with 2-D autocorrelation function

R(m,n; m,n)= E[ x(m, n) x(m, n) ] for all 0 m, m, n, n N-1
K-L Basis images is the orthonormalized eigen functions of R
Rewrite images into vector form (N2x1)

Need solve the eigen problem for N2xN2 matrix! ~ O(N 6)
Reduced computation for separable R

R(m,n; m,n)= r1(m,m) r2(n,n)
Only need solve the eigen problem for two NxN matrices ~ O(N3)
KLT can now be performed separably on rows and columns
Reducing the transform complexity from O(N4) to O(N3)
Pros and Cons of K-L Transform
Optimality
Decorrelation and MMSE for the same# of partial transf coeff.
Data dependent
Have to estimate the 2nd-order statistics to determine the transform
Can we get data-independent transform with similar performance?
DCT
Applications
(non-universal) compression
Pattern recognition: e.g., eigen faces
Analyze the principal (dominating) components and reduce
feature dimensions
Energy Compaction of DCT vs. KLT
DCT has excellent energy compaction

for highly correlated data
DCT is a good replacement for K-L

Close to optimal for highly correlated data
Not depend on specific data like K-L does
Fast algorithm available
[ref and statistics: Jains pp153, 168-175]
Energy Compaction of DCT vs. KLT (contd)

Preliminaries
The matrices R, R-1, and R-1 share the same set of eigen vectors
DCT basis vectors are eigenvectors of a symmetric tri-diagonal matrix Qc
Covariance matrix R of 1st-order stationary Markov sequence with has an

inverse in the form of symmetric tri-diagonal matrix
DCT is close to KLT on 1st-order stationary Markov

For highly correlated sequence, a scaled version of R-1 approx. Qc
More on Rate-Distortion Based

Bit Allocation in Image Coding
Basic Steps in R-D Optimization
Determine sources statistics

Simplified models: Gaussian, Laplacian, Generalized Gaussian
Obtain from training samples
Obtain operational R-D points/functions/curves

E.g., compute distortion for each candidate quantizer
Determine objective function and constraints
Search for optimal operational R-D points

Lagrange multiplier approach
select only operating points on the convex hull of overall R-D

characteristics (Fig.19 of Ortegas tutorial)
may handle different coding unit independently
Dynamic programming approach
not constrained by convex hull but has larger search space

~ higher complexity => See Ortegas SPM98 tutorial for details
Bridging the Theory and the Practice
Operational R-D curves

Directly achievable R-D pairs with practical implementation
Tradeoff!
How close are the R-D operational points to info. theory bounds?
Practical considerations: cost of memory, computation, delay, etc.
Often narrowing down to an optimization problem:
~ optimize an objective function subject to a set of constraints
Problem of model mismatch

How good are our assumptions on source modeling?
How well does a coding algorithm tolerate model mismatch?
Ref: convex hull
Formally
Smallest convex set containing all points
For R-D problem, we concerned of the lower boundary of this set that
close to the info. theoretical bound (the upper/right end are open)
Convex set
if X1,..,Xn are in the set, then the linear combinations of them (with
non-negative combination factors) are still in the set
Intuitively
Use a rubber band to wrap the outermost points so that the band
encloses all points
Bit Allocation
How many bits to be allocated for each coeff.?

Determined by the variance of coeff.
More bits for high variance k2 to keep total MSE small
1
min D
N
E v(k ) v (k )
N 1
k 0
1
subject to
N
N 1
n
k 0
N 1
k f (nk )
k 0
B bits/coeff.
where k2 -- variance of k-th coeff. v(k); nk -- # bits allocated for v(k)

f(-) quantization distortion func.
Inverse Water-filling (from info. theory)

Try to keep same amount of error in each freq. band
See Jains pp501 for detailed bit-allocation algorithm
Details on Reverse Water-filling Solution (contd)

2
i
min 12 ln
, subj. to Di D
Di
i 1
n
Construct func. using

Lagrange multiplier
Necessary condition
Keep the same
marginal gain
2
n
i
J ( D) 12 ln Di
Di
i 1
i 1
J
1 1
0 for all i Di and Di D

Di
2 Di
n
1 i2
min max ln
,0 , subj. to Di D
i 1
2 Di
n
Necessary condition
for choosing
2
2
0
,
if
D
,
if
i
i
i
s.t. Di D
2
i
2
2
Di 0, if Di i
i , if i
Lagrangian Opt. for Indep. Budget Constraint
Previous: fix distortion, minimize total rate
Alternative: fix total rate (bit budget), minimize distortion
(Discrete) Lagrangian optimization on general source
min
d
i 1
i ,q (i )
ri , q ( i )
Constant slope optimization (e.g. in Box 6 Fig. 17 of Ortegas tutorial)

Need to determine the quantizer q(i) for each coding unit i
Lagrangian cost for each coding unit d i ,q ( i ) ri ,q ( i )
use a line with slope - to intersect with each operating point
(Fig.14)
For a given operating quality , the minimum can be computed

independently for each coding unit
min
d
i 1
i ,q (i )
ri , q ( i ) min d i , q ( i ) ri , q ( i )
i 1
=> Find operating quality satisfying the rate constraint

(from Fig. 14 of Ortegas tutorial)

(from Box 6 Fig. 17 of Ortegas tutorial)

Bridging the Theory and Ad-hoc Practice
Operational R-D curves

Directly achievable with practical implementation
Tradeoff! Tradeoff! Tradeoff!

Rate vs. Distortion
how close are the operational points to info. theory bounds?
Also tradeoff among practical considerations
Cost of memory, computation, delay, etc.
Often narrowing down to an optimization problem:

~ optimize an objective func. subj. to a set of constraints
Model mismatch problem

How good are our assumptions on source modeling?
How well does a coding algorithm tolerate model mismatch?
Basic Steps in R-D Optimization
Determine sources statistics

Simplified models: Gaussian, Laplacian, Generalized Gaussian
Obtain from training samples
Obtain operational R-D points/functions/curves

E.g., compute distortion for each candidate quantizer
Determine objective func. and constraints
Search for optimal operational R-D points

Lagrange multiplier approach
select only operating points on the convex hull of overall R-D

characteristics (Fig.19 of Ortegas tutorial)
may handle different coding unit independently
Dynamic programming approach
not constrained by convex hull but has larger search space (higher
complexity)

Wavelet Coding and Related Issues

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Wavelet Coding and Related Issues

Uploaded by

Copyright:

Available Formats

ENEE631 Spring09

Wavelet Coding and Related Issues

UMCP ENEE631 Slides (created by M.Wu 2004)

Overview and Logistics

Subband decomposition; basic ideas of EZW encoding

Exploit the structures between coefficients for removing redundancy

Assignment#3 is Due Friday noon March 13

Project#1 Posted. Due Monday April 6

Lec13 Wavelet Coding, etc. [2]

UMCP ENEE631 Slides (created by M.Wu 2004)

Review: 2-D Subband/DWT

From Usevitchs article in

Lec13 Wavelet Coding, etc. [3]

UMCP ENEE631 Slides (created by M.Wu 2001)

Review: Key Concepts in EZW

Exploit multi-resolution and self-similarity

Parent-children relation among coeff.

Coding significance map via zero-tree

Avoid large overhead by coding only significant

Successive approximation quantization

M. Wu: ENEE631 Digital Image Processing (Spring'09)

From Usevitch (IEEE

Recall: EZW Algorithm and Example

Initial threshold ~ 2 ^ floor(log2 xmax)

Dominant Pass (zig-zag across bands)

Move to subordinate list

Fig. From Usevitch (IEEE Sig. Proc. Mag. 9/01)

Output one bit for subordinate list

According to position in up/low half of quantization interval

Repeat with half threshold, until bit budget achieved

Lec13 Wavelet Coding, etc. [5]

Divide quantization interval of [32,64) by half:

UMCP ENEE631 Slides (created by M.Wu 2001)

After 1st Pass

From Usevitch (IEEE Sig.Proc. Mag. 9/01)

M. Wu: ENEE631 Digital Image Processing (Spring'09)

Lec13 Wavelet Coding, etc. [6]

UMCP ENEE631 Slides (created by M.Wu 2001)

After 2nd Pass

From Usevitch (IEEE Sig.Proc. Mag. 9/01)

[16,24) => 20, [24, 32) => 28;

M. Wu: ENEE631 Digital Image Processing (Spring'09)

Lec13 Wavelet Coding, etc. [7]

UMCP ENEE631 Slides (created by M.Wu 2001/2004)

EZW and Beyond

Can apply DWT to entire images or larger blocks than 8x8

Symbol sequence can be entropy encoded

SPIHT (Set Partitioning in Hierarchal Trees)

EBCOT (Embedded Block Coding with Optimal Truncation)

Lec13 Wavelet Coding, etc. [8]

UMCP ENEE631 Slides (created by M.Wu 2004)

A Close Look at Wavelet Transform

M. Wu: ENEE631 Digital Image Processing (Spring'09)

Lec13 Wavelet Coding, etc. [9]

UMCP ENEE631 Slides (created by M.Wu 2001/2004)

Construction of Haar Functions

Unique decomposition of integer k (p, q)

for other x [0,1]

M. Wu: ENEE631 Digital Image Processing (Spring'09)

hk(x) = h p,q(x) for x [0,1]

Lec13 Wavelet Coding, etc. [10]

UMCP ENEE631 Slides (created by M.Wu 2001)

Real and orthogonal