0 Up votes0 Down votes

93 views50 pagesJan 18, 2012

© Attribution Non-Commercial (BY-NC)

PDF, TXT or read online from Scribd

Attribution Non-Commercial (BY-NC)

93 views

Attribution Non-Commercial (BY-NC)

- Neuromancer
- The E-Myth Revisited: Why Most Small Businesses Don't Work and
- How Not to Be Wrong: The Power of Mathematical Thinking
- Drive: The Surprising Truth About What Motivates Us
- Chaos: Making a New Science
- The Joy of x: A Guided Tour of Math, from One to Infinity
- How to Read a Person Like a Book
- Moonwalking with Einstein: The Art and Science of Remembering Everything
- The Wright Brothers
- The Other Einstein: A Novel
- The 6th Extinction
- The Housekeeper and the Professor: A Novel
- The Power of Discipline: 7 Ways it Can Change Your Life
- The 10X Rule: The Only Difference Between Success and Failure
- A Short History of Nearly Everything
- The Kiss Quotient: A Novel
- The End of Average: How We Succeed in a World That Values Sameness
- Made to Stick: Why Some Ideas Survive and Others Die
- Algorithms to Live By: The Computer Science of Human Decisions
- The Universe in a Nutshell

You are on page 1of 50

Presented by: Atanas Gotchev

Digital Image Processing II

Lecture Outline

1D wavelets and FWT 2D separable wavelets 2D nonseparable wavelets

different sampling grids

Lifting scheme

easy to construct filter banks

Analysis Synthesis

x[n]

H0

x0[n]

G0

^ x[n]

H1

x1[n]

G1

Digital Image Processing II

Fast wavelet transform enables efficient computation of DWT coefs. Iteration of the analysis FB on the low-pass channel DWT coefficients are computed recursively!

Synthesis Bank

Synthesis Bank

Complexity of FWT

Number of operations proportional to: N size of data L length of filters in the filterbank (scaling and wavelet vectors)

products of 1D wavelet and scaling functions (x,y) = (x)(y) (x,y) = (x)(y) V(x,y) = (x)(y) D(x,y) = (x)(y)

2D separable FWT

Sampling in 2D

Image is split into several groups of pixels (phases) Not as straightforward as in 1D Many ways to split an image

Separable Quincunx Hexagonal...

Quincunx Downsampling

n2

n1

Image is split into two phases (cosets) Simplest nonseparable sampling scheme

Subsampling Matrix

n2

(1,1)

Basis vectors form the unit cell Subsampling matrix (dilation matrix) defines the sampling operation

n1

(1,-1)

1 1 D= 1 1

Subsampling Matrix

Defines the sampling grid For a 2D grid, D is a 2x2 matrix. There are M = |det(D)| image phases and also M samples in the unit cell. For the quincunx case, M = 2.

Quincunx PR FB needs M = 2 channels.

Digital Image Processing II

2D Subsampling Operation

D defines the sampling grid Take one coset of the image Renumber it to fit on the integer grid

xD (n1 , n2 ) = x(k1 , k2 ), k1 n1 where = D k2 n2

For the quincunx case:

1 1 D= 1 1 k1 1 1 n1 n1 + n2 k = 1 1 n = n n 2 2 1 2 xD (n1 , n2 ) = x(n1 + n2 , n2 n1 )

Digital Image Processing II

Downsampling is actually...

reading the image along the new axes. 45 rotation for the quincunx case

n2

n2

(1,1)

(0,1)

n1 (1,-1)

(1,0) n1

move the new axes by (1,0)... to the next element of the unit cell.

n2

n2

(2,1)

(0,1)

n1 (2,-1)

(1,0) n1

e as h

Ph

as e

Counterclockwise rotation

Digital Image Processing II

Separable Sampling

n2 (0,2)

(2,0) n1

4 elements of the unit cell Image is split into 4 phases Requires 4 channels of the PR filter bank

2 0 D= 0 2

Hexagonal Sampling

n2

(1,2)

4 elements of the unit cell Image is split into 4 phases Requires 4 channels of the PR filter bank

n1

(1,-2)

1 1 D= 2 2

Voronoi cell

Voronoi cell consists of points closer to the origin... than to any other point of the given lattice. n Quincunx Voronoi cell

2

1 1 n1

Downsampling is defined with a D matrix

1 X D () = ( D) X () = X D T ( 2 k ) T | det D | kN ( D )

where

1 = 2

To avoid aliasing... signal should be bandlimited to Voronoi cell of the lattice defined by 2D-T

Digital Image Processing II

Bandlimiting

Properly bandlimited signal for quincunx downsampling

2

Quincunx downsampling

Input image has been properly bandlimited

2

2 2

2

2

Quincunx upsampling

x(D1n) xU (n) = 0

n2 (0,1)

n2

(1,1)

(1,0) n1

n1 (1,-1)

X U ( z ) = xU (n)z n = x( D 1n)z n = x(k )z Dk = X ( z D )

n n k

z1 z = z2

n

n1 n2

n = z1n1 z 2 2

z1 D z = z2

d11 d 21

d12 d 22

d z1d11 z 2 21 = d d z1 12 z 2 22

Digital Image Processing II

Frequency transformation

z e j

d z1d11 z 2 21 e j ( d111 + d 212 ) D jD T z = d d j (d +d ) = e z1 12 z 2 22 e 12 1 22 2

Conclusion:

X U () = X ( D T )

Quincunx upsampling

X ()

2

2 2

X U () = X (DT )

2

2

X U () = X (DT )

2

X U () = X (D2 )T

2

)

X U () = X (D3 )T

2

Simple way to construct filter banks Easy to satisfy PR requirement Computationally efficient

X(z) z-1 2 2 + P(z) U(z) D(z) A(z) U(z) P(z) + 2 z-1 ^ X(z) 2

Basic structure:

Polyphase decomposition Predict stage (dual lifting step) Update stage (primal lifting step)

X(z) z-1 2 2 + P(z) U(z) D(z) A(z) U(z) P(z) + 2 z-1 ^ X(z) 2

Predict stage

Prediction of the second phase sample ...based on a number of samples from the first phase. Wavelet coefficients are obtained as... a prediction error. X(z) 2 Smooth signal... gives small details.

Digital Image Processing II

z-1 2

P(z) D(z)

Update stage

Input: detail coefs. Output is used to create approximation coefs. Average value of the input image must be retained. X(z) A(z)

2 + z-1 2

Digital Image Processing II

P(z) -

U(z) D(z)

X(z1,z2) D Xe + P(z1,z2) z1-1 D Xo U(z1,z2) D A U(z1,z2) P(z1,z2) + D z1 ^ X(z1,z2) D

Digital Image Processing II

Quincunx FB Example

Lifting scheme based on quincunx interpolating filters J. Kovaevi & W. Sweldens: Wavelet Families of Increasing Order in Arbitrary Dimensions. IEEE Trans. Image Proc., vol. 9, no. 3, pages 480-496, March 2000.

Predict Filters

Neville interpolating filters symmetric interpolation neighborhoods

n2

n1

example of a second order P filter: P2 ( z1 , z 2 ) = 0.25 + 0.25 z11 + 0.25 z 2 1 + 0.25 z11 z 2 1

Digital Image Processing II

Update Filters

updates the average value of the input image based on the corresponding predict filter

1 * U N ( z1 , z2 ) = PN ( z1 , z2 ) 2

Digital Image Processing II

Analysis LP

Analysis HP

Synthesis LP

Digital Image Processing II

Synthesis HP

Analysis scale

Analysis wavelet

Synthesis scale

Digital Image Processing II

Synthesis wavelet

AJ-1 AJ-2 AJ-3

DJ-1

DJ-2

DJ-3

Nonseparable

higher complexity more freedom in FB design different directional properties

Separable

widely used simple realization based on 1D filter banks

Quincunx Wavelets

Simplest nonseparable sampling grid Only two channels Double quincunx sampling = nonseparable sampling Less biased in horizontal and vertical directions Comparable results with separable wavelets

Digital Image Processing II

- MAN231 SensitivitasUploaded bynikeprimarini
- 3D_waveletTransUploaded bysalilpn2
- Final ReportUploaded byRahul Yadav
- Bharath Ms ThesisUploaded bygsajer
- Erg MeanUploaded byÇháråñ Çhèrry
- 1-s2.0-S1474667015388637-mainUploaded byanon_509549634
- li-2008-icasspUploaded byHunter Vn
- Applied MathematicsUploaded bysgxrgsys
- Optimization of Gas-Injected Oil WellsUploaded bySufian R Ellabbad
- POTWE-13-CP-03-SUploaded byscribd-in-action
- 741-T017.pdfUploaded bymkmanish1
- MITRES_6_007S11_hw19Uploaded byxabihdez05
- WAVELETS AND PARITY EQUATIONS METHODS BASED POSITION SENSOR FAULT DETECTION IN PMSM DRIVES.pdfUploaded bysupernoufa
- Loop CardsUploaded byJulia Smith
- BfsUploaded byraden hilmi
- Mathmatical ModelUploaded byVinod Thete
- Transportation ModelUploaded bySamantha Weinert
- Credit CalcUploaded byveda20
- bachPaopy2Uploaded bykaysheph
- CWUploaded byGerald Thompson
- wqqwUploaded bySaagar Karande
- dfsimplementationUploaded bySaurabh Middha
- Bug List AUploaded bybarneymartyn
- Assignment 1-3-2011Uploaded bySoumendu Majee
- Tute Sheet 05 UMA004Uploaded byJapanjOt SinGh
- Math 1313 Printable Version Quiz 9Uploaded byMcenroe
- Enhancing Spectrum Utilization using GA in WCDMAUploaded byIJRASETPublications
- Match Fltr Using PNUploaded bykenjo138
- Unit CommitmentUploaded bysusmita000
- iitUploaded bylucas

- Compiler ConstructionUploaded bykykol
- MonteCarloUploaded byHưng Nguyễn Ngọc
- book armUploaded byKleber Freitas
- Ps3000ab ProgrammerGuide.enUploaded byHưng Nguyễn Ngọc
- Software EngineeringUploaded byHưng Nguyễn Ngọc
- Fundamentals_of_Building_a_Test_System_CompleteGuide.pdfUploaded byHưng Nguyễn Ngọc
- 6 Essential Tips of Using Osciloscope AgilentUploaded byHưng Nguyễn Ngọc
- 00.LaplaceTransformTableUploaded byHưng Nguyễn Ngọc
- CUDA C Programming GuideUploaded byReyhan Kara
- RGB_to_HSIUploaded byVaishali Swamy

- BE CollegesUploaded byRohan Joshi
- __4047__Uploaded byrayan
- Arbiarso Wijatmoko in 2018Uploaded byArbiarso Wijatmoko
- AquaUploaded byPolipio Saturnio
- Influence of P-Delta effect on reinforced concrete buildings with vertical irregularity - A ReviewUploaded byAnonymous CUPykm6DZ
- Phasing in DCS Migration[1]Uploaded bypanchogancho
- PopQuizCh3Au11Uploaded byChristian De Leon
- Applying Baldridge Organizational Effectiveness Model for Accreditation of Higher Education InstitutionUploaded byMuhammad Dicky
- Oracle Database 11g on Windows Dev 134379Uploaded byp_palmieri
- 5. Sca_beltran v.phhcUploaded byClarkBarcelon
- Active DirectoryUploaded byAnonymous NeRBrZyAUb
- Sales and distribution Report On P&G's Hair Care Product in Cuttack CityUploaded byShovan Dash
- 09_clx_sentinel_smartUploaded bymoazedi_moj
- Econ Q2(2HH03)SolUploaded byMay
- BRIEF REVIEW OF SOLID ELECTROLYTE FOR LITHIUM ION BATTERIES IN PARTICULAR TO GARNET-STRUCTURED Li7La3Zr2O12 SOLID-STATE ELECTROLYTE.Uploaded byIJAR Journal
- Quality Inspection and Testing of Household Sewing Machines HeadUploaded byHarshaDesu
- ITB and Terms of Reference for Detailed Architectural and Engineering Design for the Construction of UPOU Learning CommonsUploaded byUpou Spmo
- Black HoleUploaded bysharmasunil6325
- Scrip Cryptotab Hack 2018 _ Computer Related Introductions _ TextUploaded byGizli Bilgiler
- Microsoft PowerPoint - 5-OWF906206 Subscriber Data ManagemenUploaded byGeni Evans
- Developing an International Financial Center to Modernize the Korean Service Sector by Yoon-shik ParkUploaded byKorea Economic Institute of America (KEI)
- Itc Vishal FinalUploaded byVishal Zalavadiya
- ABSTRACT Ratio AnalysisUploaded byDiwakar Srivastava
- Oil & Gas Exploration & Production FundamentalsUploaded bymicktunx
- MESSA 2008 AuditUploaded byR1Lawrence
- Truss DetailUploaded byVincent Pilien
- Software Development FundamentalsUploaded byNathanDrake
- Everything You Need to Know About TX ValvesUploaded bycalishe1
- CHM3960 2011 ExamUploaded bypdoleary
- Cdvdt SolutionUploaded byKiran Kumar

## Much more than documents.

Discover everything Scribd has to offer, including books and audiobooks from major publishers.

Cancel anytime.