Midterm Summary Fall 08. Yao Wang Polytechnic University, Brooklyn, NY 11201

Size: px
Start display at page:

Download "Midterm Summary Fall 08. Yao Wang Polytechnic University, Brooklyn, NY 11201"

Transcription

1 Midterm Summary Fall 8 Yao Wang Polytechnic University, Brooklyn, NY 2

2 Components in Digital Image Processing Output are images Input Image Color Color image image processing Image Image restoration Image Image enhancement Image Image acquisition Wavelets and and Multiresolution processing Knowledge base base Compression Morphological processing Segmentation Representation & description Object Object recognition Output are image attributes Yao Wang, NYU-Poly EL523: Midterm Summary 2

3 Topics Covered Image representation Color representation Quantization Contrast enhancement Spatial Filtering: noise removal, sharpening, edge detection Frequency domain representations FT, DTFT, DFT Implementation o linear iltering using DTFT and DFT Yao Wang, NYU-Poly EL523: Midterm Summary 3

4 Color Image Representation Light is the visible band o the EM wave Color hue o a light depends on the wavelength o the EM wave Illuminating light vs. relecting light Color attributes: Brightness luminance Hue depends on the wavelength Saturation purity Any color can be reproduced by mixing three primary colors Illuminating and relecting light sources ollow dierent mixing rules Color can be represented in many dierent coordinates or models with 3 components Yao Wang, NYU-Poly EL523: Midterm Summary 4

5 Tri-chromatic Color Mixing Tri-chromatic color mixing theory Any color can be obtained by mixing three primary colors with a right proportion Primary colors or illuminating sources: Red, Green, Blue RGB Color monitor works by exciting red, green, blue phosphors using separate electronic guns ollows additive rule: R+G+BWhite Primary colors or relecting sources also known as secondary colors: Cyan, Magenta, Yellow CMY Color printer works by using cyan, magenta, yellow and black CMYK dyes ollows subtractive rule: R+G+BBlack Yao Wang, NYU-Poly EL523: Midterm Summary 5

6 Color Models Speciy three primary or secondary colors Red, Green, Blue. Cyan, Magenta, Yellow. Speciy the luminance and chrominance HSB or HSI Hue, saturation, and brightness or intensity YIQ used in NTSC color TV YCbCr used in digital color TV Amplitude speciication: 8 bits per color component, or 24 bits per pixel Total o 6 million colors A kxk true RGB color requires 3 MB memory Yao Wang, NYU-Poly EL523: Midterm Summary 6

7 Color Image Processing Apply contrast enhancement or iltering linear or non-linear to each primary color component independently using the techniques or monochrome images May change the color hue o the original image Convert the tri-stimulus representation into a luminance / chrominance representation, and modiy the luminance component only. For certain applications, dierent operations may be applied to dierent color components either in RGB domain or in YCC2 domain to obtain special desired eects Yao Wang, NYU-Poly EL523: Midterm Summary 7

8 Pseudo Color Processing Given a gray-scale image, we may display it in pseudo colors to better reveal certain eatures Color depends on image pixel value Oten used in medical image display, weather maps, vegetation maps Decompose an image into dierent requency bands and represent each band with dierent colors Yao Wang, NYU-Poly EL523: Midterm Summary 8

9 Quantization Q Quantizer Q r 8 Decision Levels {t k, k,, L+} Reconstruction Levels {r k, k,, L} I t k, t [ k + Then Q r k t t 2 t 3 t 4 t 5 t 6 t 7 t 8 L levels need bits x R log2 returns the smallest integer that is bigger than or equal to x L r 7 r 6 r 5 r 4 r 3 r 2 r Quantizer error r r r 2 r 3 r 4 r 5 r 6 r 7 r 8 r t t 2 t 3 t 4 t 5 t 6 t 7 t 8 Yao Wang, NYU-Poly EL523: Midterm Summary 9

10 Uniorm Quantization Equal distances between adjacent decision levels and between adjacent reconstruction levels t l t l- r l r l- q Parameters o Uniorm Quantization L: levels L 2 R B: dynamic range B max min q: quantization interval step size q B/L B2 -R Yao Wang, NYU-Poly EL523: Midterm Summary

11 Uniorm Quantization: Functional Representation Q stepsize q max - min /L r 7 max -q/2 r 6 q Q + q 2 min q + min r 5 r 3 r 4 x returns the biggest integer that is smaller than or equal to x r 2 r r min +q/2 t t t 2 t 3 t 4 t 5 t 6 t 7 t 8 min max min I is called the reconstruction level index, which indicates which reconstruction level is used or. q Yao Wang, NYU-Poly EL523: Midterm Summary

12 Contrast Enhancement An requently used, important operation How to tell the contrast o an image rom its histogram? Given a histogram, can sketch a transormation that will enhance the contrast Histogram equalization Histogram speciication Yao Wang, NYU-Poly EL523: Midterm Summary 2

13 Histogram vs. Contrast p p p a Too dark b Too bright c Well balanced Yao Wang, NYU-Poly EL523: Midterm Summary 3

14 Enhancement o Too-Dark Images H g g max New histogram max Hg Original histogram max Transormation unction: log unction: gc log + Or Power law: g c ^r, <r< g max g Yao Wang, NYU-Poly EL523: Midterm Summary 4

15 Enhancement o Too-Bright Images H g g max New histogram max Hg Original histogram max Transormation unction: Power law: gc ^r, r> g max g Yao Wang, NYU-Poly EL523: Midterm Summary 5

16 Enhancement o Images Centered near the Middle Range H g g max New histogram max Hg Original histogram max Transormation unction g max g Yao Wang, NYU-Poly EL523: Midterm Summary 6

17 Histogram Equalization Transorms an image with an arbitrary histogram to one with a lat histogram Suppose has PDF p F, Transorm unction continuous version g g is uniormly distributed in, pf t dt Histogram Equalization Yao Wang, NYU-Poly EL523: Midterm Summary 7

18 Histogram Speciication What i the desired histogram is not lat? p F Histogram Equalization g s Histogram Equalization z p Z z p G g g g P S s s s z Speciied Histogram p F d z p Z z dz z s g s g Yao Wang, NYU-Poly EL523: Midterm Summary 8

19 Image Filtering Applications: Noise removal image smoothing, low-pass iltering Edge enhancement deblurring, high-emphasis Edge detection high-pass iltering Contrast enhancement is accomplished by point operation, i.e., each pixel value is changed based on its original value, not its neighboring pixels but the transormation unction depends on the overall histogram o the image Image iltering reers to changing the color value o one pixel based on the color values o this pixel and its neighbors Yao Wang, NYU-Poly EL523: Midterm Summary 9

20 Three Ways o Implementing Linear Filtering Spatial domain Weighted average o adjacent pixels linear convolution Weights or the ilter depends on the desired iltering eect Frequency domain FT, DTFT Design spatial ilter based on desired requency response Low pass, high pass, high emphasis Convolution theorem: h* H F Frequency domain DFT Perorm iltering in DFT domain Convolution theorem: circulation convolution h@ H F Relation between circulation convolution and linear convolution Filter masks in the DFT domain must be designed properly so that the corresponding ilter in the spatial domain is real The ilter mask should enjoy the symmetry property o real signals Yao Wang, NYU-Poly EL523: Midterm Summary 2

21 Yao Wang, NYU-Poly EL523: Midterm Summary 2 Linear Convolution o Continuous Signals D convolution Equalities 2D convolution α α α α α α d x h d h x x h x, x x x x x x x δ δ β α β α β α β α β α β α d d y x h d d h y x y x h y x,,,,,,

22 Yao Wang, NYU-Poly EL523: Midterm Summary 22 Linear Convolution o Discrete Signals D convolution 2D convolution Separable iltering Row irst, then columns m m m n h m m h m n n h n * k l k l l n k m h l k l k h l n k m n m h n m,,,,, *,

23 Interpretation as weighted average o neighboring samples How should we design the weights? Smoothing: sum, coe > High emphasis: sum, some coe < Edge detection high pass: sum Yao Wang, NYU-Poly EL523: Midterm Summary 23

24 Ex: Smoothing by Averaging Replace each pixel by the average o pixels in a square window surrounding this pixel g m, n m, n + m, n + m +, n + m, n + m, n + m, n + m +, n + m, n + m +, n + Trade-o between noise removal and detail preserving: Larger window -> can remove noise more eectively, but also blur the details/edges Yao Wang, NYU-Poly EL523: Midterm Summary 24

25 Yao Wang, NYU-Poly EL523: Midterm Summary 25 Directional Edge Detector High pass in one direction and low pass in the orthogonal direction Prewitt edge detector Sobel edge detector [ ] [ ] 3 3 ; 3 3 y H x H [ ] [ ] ; y H x H

26 Special Considerations in Implementation Boundary treatment I is MxN, h is KxL, convolved image g is M+K- xn+l- Instead o expanding the image size, we modiy iltering operations at the boundary Symmetric expansion Leave the boundary pixels unchanged Renormalization Filtered values may not be integer, and may have negative values, may have a smaller or larger dynamic range than original To save resulting image in an 8-bit unsigned char ormat, we normalize all values to -255 g g-g_min/g_max-g_min * 255 Yao Wang, NYU-Poly EL523: Midterm Summary 26

27 D Fourier Transorm For Discrete Time Sequence DTFT Forward Transorm Inverse Transorm j2πun n e n / 2 j2πun n F u e du / 2 F u 2D Forward Transorm F u, v m n m, n e j2π mu+ nv Inverse Transorm / 2 / 2 j2π mu+ nv m, n F u, v e dudv / 2 / 2 Separable implementation Transorm each row irst with D DTFT, then each column Yao Wang, NYU-Poly EL523: Midterm Summary 27

28 Design Filters Based on Desired Frequency Response Convolution theorem x, y* h x, y F u, v H u, v Filter hx,y designed based on desired requency response Hu,v Sharp transitions in requency domain very long ilters in spatial domain Apply a window unction to smooth the transition band shorter ilters Giver a spatial ilter, we can use DTFT to better understand its iltering eect requency response Separable ilters Design horizontal and vertical ilters separately For images, we usually use very short ilters Yao Wang, NYU-Poly EL523: Midterm Summary 28

29 Discrete Fourier Transorm DFT: I DTFT or Finite Duration Signals the signal is only deined or n Fourier transorm becomes : F' N n nexp,,..., N j2πn,, : Sampling F' at Forward transorm DFT : N k F k F' N N n Inverse transorm IDFT : n N N k k/n, k k F kexp j2π N,,...,N-, and rescaling yields: nexp n, k j2π N n, n,,..., N k,,..., N Yao Wang, NYU-Poly EL523: Midterm Summary 29

30 Circular Convolution Theorem Circular convolution n h n N k n k N h k n-k N n hn n n n n- Convolution theorem n h n F k H k n+ n+2 n+4 n+3 k Yao Wang, NYU-Poly EL523: Midterm Summary 3

31 2D Discrete Fourier Transorm Deinition Assuming m, n, m,,, M-, n,,, N-, is a inite length 2D sequence F k, l m, n MN MN M N m n M N k l m, n e F k, l e Comparing to DTFT F u, v m n m, n e km ln j2π + M N km ln j2π + M N,, j2π mu+ nv k,,..., M m,,..., M, u k M, l, n, v / 2 / 2 j2π mu+ nv m, n F u, v e dudv / 2 / 2 l N,,..., N,,..., N ;. Yao Wang, NYU-Poly EL523: Midterm Summary 3

32 Periodicity Property o 2D DFT Periodicity F k, l F k, l, k < or k M, l < or l N. M N -M -M/2 A C B D A C B D Low Frequency k M/2 M -N A C B D,,N- A C M-, B D M-,N- -N/2 N/2 N l High Frequency Yao Wang, NYU-Poly EL523: Midterm Summary 32

33 Calculate Linear Convolution Using DFT D case n is length N, hn is length N 2 gn n*hh is length N N +N 2 -. To use DFT, need to extend n and hn to length N by zero padding. n * hn Convolution gn DFT DFT DFT Fk x Hk Multiplication Gk Yao Wang, NYU-Poly EL523: Midterm Summary 33

34 Midterm Exam Logistics Scheduled time: /5 3-5:4, LC2 Closed-book, sheet o notes allowed double sided OK Oice hour by the Grader Zhi Zhao, Tuesday /4-2AM Wed. /5-2 AM LC22 Zhi Zhao [zhaozhi429@yahoo.com.cn] Yao Wang, NYU-Poly EL523: Midterm Summary 34

Midterm Summary Fall Yao Wang Polytechnic University, Brooklyn, NY 11201

Midterm Summary Fall Yao Wang Polytechnic University, Brooklyn, NY 11201 Midterm Summary Fa 3 Yao Wang Poytecnic University Brooyn NY Topics Covered Image representation Coor representation Quantization Contrast enancement Spatia Fitering: noise remova sarpening edge detection

More information

Department of Electrical Engineering, Polytechnic University, Brooklyn Fall 05 EL DIGITAL IMAGE PROCESSING (I) Final Exam 1/5/06, 1PM-4PM

Department of Electrical Engineering, Polytechnic University, Brooklyn Fall 05 EL DIGITAL IMAGE PROCESSING (I) Final Exam 1/5/06, 1PM-4PM Department of Electrical Engineering, Polytechnic University, Brooklyn Fall 05 EL512 --- DIGITAL IMAGE PROCESSING (I) Y. Wang Final Exam 1/5/06, 1PM-4PM Your Name: ID Number: Closed book. One sheet of

More information

Digital Image Processing. Lecture 6 (Enhancement) Bu-Ali Sina University Computer Engineering Dep. Fall 2009

Digital Image Processing. Lecture 6 (Enhancement) Bu-Ali Sina University Computer Engineering Dep. Fall 2009 Digital Image Processing Lecture 6 (Enhancement) Bu-Ali Sina University Computer Engineering Dep. Fall 009 Outline Image Enhancement in Spatial Domain Spatial Filtering Smoothing Filters Median Filter

More information

CAP 5415 Computer Vision Fall 2011

CAP 5415 Computer Vision Fall 2011 CAP 545 Computer Vision Fall 2 Dr. Mubarak Sa Univ. o Central Florida www.cs.uc.edu/~vision/courses/cap545/all22 Oice 247-F HEC Filtering Lecture-2 General Binary Gray Scale Color Binary Images Y Row X

More information

Lecture 13: Applications of Fourier transforms (Recipes, Chapter 13)

Lecture 13: Applications of Fourier transforms (Recipes, Chapter 13) Lecture 13: Applications o Fourier transorms (Recipes, Chapter 13 There are many applications o FT, some o which involve the convolution theorem (Recipes 13.1: The convolution o h(t and r(t is deined by

More information

Lecture Outline. Basics of Spatial Filtering Smoothing Spatial Filters. Sharpening Spatial Filters

Lecture Outline. Basics of Spatial Filtering Smoothing Spatial Filters. Sharpening Spatial Filters 1 Lecture Outline Basics o Spatial Filtering Smoothing Spatial Filters Averaging ilters Order-Statistics ilters Sharpening Spatial Filters Laplacian ilters High-boost ilters Gradient Masks Combining Spatial

More information

CAP 5415 Computer Vision

CAP 5415 Computer Vision CAP 545 Computer Vision Dr. Mubarak Sa Univ. o Central Florida Filtering Lecture-2 Contents Filtering/Smooting/Removing Noise Convolution/Correlation Image Derivatives Histogram Some Matlab Functions General

More information

Image Acquisition and Sampling Theory

Image Acquisition and Sampling Theory Image Acquisition and Sampling Theory Electromagnetic Spectrum The wavelength required to see an object must be the same size of smaller than the object 2 Image Sensors 3 Sensor Strips 4 Digital Image

More information

Optimum Notch Filtering - I

Optimum Notch Filtering - I Notch Filtering 1 Optimum Notch Filtering - I Methods discussed before: filter too much image information. Optimum: minimizes local variances of the restored image fˆ x First Step: Extract principal freq.

More information

Digital Image Processing COSC 6380/4393

Digital Image Processing COSC 6380/4393 Digital Image Processing COSC 6380/4393 Lecture 11 Oct 3 rd, 2017 Pranav Mantini Slides from Dr. Shishir K Shah, and Frank Liu Review: 2D Discrete Fourier Transform If I is an image of size N then Sin

More information

Introduction to Analog And Digital Communications

Introduction to Analog And Digital Communications Introduction to Analog And Digital Communications Second Edition Simon Haykin, Michael Moher Chapter Fourier Representation o Signals and Systems.1 The Fourier Transorm. Properties o the Fourier Transorm.3

More information

Empirical Mean and Variance!

Empirical Mean and Variance! Global Image Properties! Global image properties refer to an image as a whole rather than components. Computation of global image properties is often required for image enhancement, preceding image analysis.!

More information

Notes on Wavelets- Sandra Chapman (MPAGS: Time series analysis) # $ ( ) = G f. y t

Notes on Wavelets- Sandra Chapman (MPAGS: Time series analysis) # $ ( ) = G f. y t Wavelets Recall: we can choose! t ) as basis on which we expand, ie: ) = y t ) = G! t ) y t! may be orthogonal chosen or appropriate properties. This is equivalent to the transorm: ) = G y t )!,t )d 2

More information

Image Filtering. Slides, adapted from. Steve Seitz and Rick Szeliski, U.Washington

Image Filtering. Slides, adapted from. Steve Seitz and Rick Szeliski, U.Washington Image Filtering Slides, adapted from Steve Seitz and Rick Szeliski, U.Washington The power of blur All is Vanity by Charles Allen Gillbert (1873-1929) Harmon LD & JuleszB (1973) The recognition of faces.

More information

Introduction to Computer Vision. 2D Linear Systems

Introduction to Computer Vision. 2D Linear Systems Introduction to Computer Vision D Linear Systems Review: Linear Systems We define a system as a unit that converts an input function into an output function Independent variable System operator or Transfer

More information

Lab 3: The FFT and Digital Filtering. Slides prepared by: Chun-Te (Randy) Chu

Lab 3: The FFT and Digital Filtering. Slides prepared by: Chun-Te (Randy) Chu Lab 3: The FFT and Digital Filtering Slides prepared by: Chun-Te (Randy) Chu Lab 3: The FFT and Digital Filtering Assignment 1 Assignment 2 Assignment 3 Assignment 4 Assignment 5 What you will learn in

More information

Probabilistic Model of Error in Fixed-Point Arithmetic Gaussian Pyramid

Probabilistic Model of Error in Fixed-Point Arithmetic Gaussian Pyramid Probabilistic Model o Error in Fixed-Point Arithmetic Gaussian Pyramid Antoine Méler John A. Ruiz-Hernandez James L. Crowley INRIA Grenoble - Rhône-Alpes 655 avenue de l Europe 38 334 Saint Ismier Cedex

More information

Wavelets and Multiresolution Processing

Wavelets and Multiresolution Processing Wavelets and Multiresolution Processing Wavelets Fourier transform has it basis functions in sinusoids Wavelets based on small waves of varying frequency and limited duration In addition to frequency,

More information

Image Enhancement in the frequency domain. GZ Chapter 4

Image Enhancement in the frequency domain. GZ Chapter 4 Image Enhancement in the frequency domain GZ Chapter 4 Contents In this lecture we will look at image enhancement in the frequency domain The Fourier series & the Fourier transform Image Processing in

More information

EDGES AND CONTOURS(1)

EDGES AND CONTOURS(1) KOM31 Image Processing in Industrial Sstems Dr Muharrem Mercimek 1 EDGES AND CONTOURS1) KOM31 Image Processing in Industrial Sstems Some o the contents are adopted rom R. C. Gonzalez, R. E. Woods, Digital

More information

Chapter 4 Image Enhancement in the Frequency Domain

Chapter 4 Image Enhancement in the Frequency Domain Chapter 4 Image Enhancement in the Frequency Domain 3. Fourier transorm -D Let be a unction o real variable,the ourier transorm o is F { } F u ep jπu d j F { F u } F u ep[ jπ u ] du F u R u + ji u or F

More information

Color Science Light & Spectra

Color Science Light & Spectra COLOR Visible Light Color Science Light & Spectra Light is an electromagnetic wave It s color is characterized by it s wavelength Most light sources produce contributions over many wavelengths, contributions

More information

Chapter 4 Image Enhancement in the Frequency Domain

Chapter 4 Image Enhancement in the Frequency Domain Chapter 4 Image Enhancement in the Frequency Domain Yinghua He School of Computer Science and Technology Tianjin University Background Introduction to the Fourier Transform and the Frequency Domain Smoothing

More information

Multidimensional Signal Processing

Multidimensional Signal Processing Multidimensional Signal Processing Mark Eisen, Alec Koppel, and Alejandro Ribeiro Dept. of Electrical and Systems Engineering University of Pennsylvania aribeiro@seas.upenn.edu http://www.seas.upenn.edu/users/~aribeiro/

More information

Filtering in Frequency Domain

Filtering in Frequency Domain Dr. Praveen Sankaran Department of ECE NIT Calicut February 4, 2013 Outline 1 2D DFT - Review 2 2D Sampling 2D DFT - Review 2D Impulse Train s [t, z] = m= n= δ [t m T, z n Z] (1) f (t, z) s [t, z] sampled

More information

EEO 401 Digital Signal Processing Prof. Mark Fowler

EEO 401 Digital Signal Processing Prof. Mark Fowler EEO 401 Digital Signal Processing Pro. Mark Fowler Note Set #14 Practical A-to-D Converters and D-to-A Converters Reading Assignment: Sect. 6.3 o Proakis & Manolakis 1/19 The irst step was to see that

More information

8 The Discrete Fourier Transform (DFT)

8 The Discrete Fourier Transform (DFT) 8 The Discrete Fourier Transform (DFT) ² Discrete-Time Fourier Transform and Z-transform are de ned over in niteduration sequence. Both transforms are functions of continuous variables (ω and z). For nite-duration

More information

Computer Vision & Digital Image Processing

Computer Vision & Digital Image Processing Computer Vision & Digital Image Processing Image Restoration and Reconstruction I Dr. D. J. Jackson Lecture 11-1 Image restoration Restoration is an objective process that attempts to recover an image

More information

Multimedia Information Systems

Multimedia Information Systems Multimedia Information Systems Samson Cheung EE 639, Fall 2004 Lecture 3 & 4: Color, Video, and Fundamentals of Data Compression 1 Color Science Light is an electromagnetic wave. Its color is characterized

More information

Review Smoothing Spatial Filters Sharpening Spatial Filters. Spatial Filtering. Dr. Praveen Sankaran. Department of ECE NIT Calicut.

Review Smoothing Spatial Filters Sharpening Spatial Filters. Spatial Filtering. Dr. Praveen Sankaran. Department of ECE NIT Calicut. Spatial Filtering Dr. Praveen Sankaran Department of ECE NIT Calicut January 7, 203 Outline 2 Linear Nonlinear 3 Spatial Domain Refers to the image plane itself. Direct manipulation of image pixels. Figure:

More information

Multiscale Image Transforms

Multiscale Image Transforms Multiscale Image Transforms Goal: Develop filter-based representations to decompose images into component parts, to extract features/structures of interest, and to attenuate noise. Motivation: extract

More information

Enhancement Using Local Histogram

Enhancement Using Local Histogram Enhancement Using Local Histogram Used to enhance details over small portions o the image. Deine a square or rectangular neighborhood hose center moves rom piel to piel. Compute local histogram based on

More information

ENSC327 Communications Systems 2: Fourier Representations. School of Engineering Science Simon Fraser University

ENSC327 Communications Systems 2: Fourier Representations. School of Engineering Science Simon Fraser University ENSC37 Communications Systems : Fourier Representations School o Engineering Science Simon Fraser University Outline Chap..5: Signal Classiications Fourier Transorm Dirac Delta Function Unit Impulse Fourier

More information

Image Enhancement (Spatial Filtering 2)

Image Enhancement (Spatial Filtering 2) Image Enhancement (Spatial Filtering ) Dr. Samir H. Abdul-Jauwad Electrical Engineering Department College o Engineering Sciences King Fahd University o Petroleum & Minerals Dhahran Saudi Arabia samara@kupm.edu.sa

More information

Compressing a 1D Discrete Signal

Compressing a 1D Discrete Signal Compressing a D Discrete Signal Divide the signal into 8blocks. Subtract the sample mean from each value. Compute the 8 8covariancematrixforthe blocks. Compute the eigenvectors of the covariance matrix.

More information

EECE 301 Signals & Systems Prof. Mark Fowler

EECE 301 Signals & Systems Prof. Mark Fowler EECE 3 Signals & Systems Pro. Mark Fowler Discussion #9 Illustrating the Errors in DFT Processing DFT or Sonar Processing Example # Illustrating The Errors in DFT Processing Illustrating the Errors in

More information

Digital Image Processing

Digital Image Processing Digital Image Processing Image Transforms Unitary Transforms and the 2D Discrete Fourier Transform DR TANIA STATHAKI READER (ASSOCIATE PROFFESOR) IN SIGNAL PROCESSING IMPERIAL COLLEGE LONDON What is this

More information

Filtering and Edge Detection

Filtering and Edge Detection Filtering and Edge Detection Local Neighborhoods Hard to tell anything from a single pixel Example: you see a reddish pixel. Is this the object s color? Illumination? Noise? The next step in order of complexity

More information

encoding without prediction) (Server) Quantization: Initial Data 0, 1, 2, Quantized Data 0, 1, 2, 3, 4, 8, 16, 32, 64, 128, 256

encoding without prediction) (Server) Quantization: Initial Data 0, 1, 2, Quantized Data 0, 1, 2, 3, 4, 8, 16, 32, 64, 128, 256 General Models for Compression / Decompression -they apply to symbols data, text, and to image but not video 1. Simplest model (Lossless ( encoding without prediction) (server) Signal Encode Transmit (client)

More information

! Circular Convolution. " Linear convolution with circular convolution. ! Discrete Fourier Transform. " Linear convolution through circular

! Circular Convolution.  Linear convolution with circular convolution. ! Discrete Fourier Transform.  Linear convolution through circular Previously ESE 531: Digital Signal Processing Lec 22: April 18, 2017 Fast Fourier Transform (con t)! Circular Convolution " Linear convolution with circular convolution! Discrete Fourier Transform " Linear

More information

IMAGE ENHANCEMENT II (CONVOLUTION)

IMAGE ENHANCEMENT II (CONVOLUTION) MOTIVATION Recorded images often exhibit problems such as: blurry noisy Image enhancement aims to improve visual quality Cosmetic processing Usually empirical techniques, with ad hoc parameters ( whatever

More information

In many diverse fields physical data is collected or analysed as Fourier components.

In many diverse fields physical data is collected or analysed as Fourier components. 1. Fourier Methods In many diverse ields physical data is collected or analysed as Fourier components. In this section we briely discuss the mathematics o Fourier series and Fourier transorms. 1. Fourier

More information

Lecture 3: Linear Filters

Lecture 3: Linear Filters Lecture 3: Linear Filters Professor Fei Fei Li Stanford Vision Lab 1 What we will learn today? Images as functions Linear systems (filters) Convolution and correlation Discrete Fourier Transform (DFT)

More information

BANDELET IMAGE APPROXIMATION AND COMPRESSION

BANDELET IMAGE APPROXIMATION AND COMPRESSION BANDELET IMAGE APPOXIMATION AND COMPESSION E. LE PENNEC AND S. MALLAT Abstract. Finding eicient geometric representations o images is a central issue to improve image compression and noise removal algorithms.

More information

Scattered Data Approximation of Noisy Data via Iterated Moving Least Squares

Scattered Data Approximation of Noisy Data via Iterated Moving Least Squares Scattered Data Approximation o Noisy Data via Iterated Moving Least Squares Gregory E. Fasshauer and Jack G. Zhang Abstract. In this paper we ocus on two methods or multivariate approximation problems

More information

Lecture 3: Linear Filters

Lecture 3: Linear Filters Lecture 3: Linear Filters Professor Fei Fei Li Stanford Vision Lab 1 What we will learn today? Images as functions Linear systems (filters) Convolution and correlation Discrete Fourier Transform (DFT)

More information

Digital Signal Processing. Lecture Notes and Exam Questions DRAFT

Digital Signal Processing. Lecture Notes and Exam Questions DRAFT Digital Signal Processing Lecture Notes and Exam Questions Convolution Sum January 31, 2006 Convolution Sum of Two Finite Sequences Consider convolution of h(n) and g(n) (M>N); y(n) = h(n), n =0... M 1

More information

Lecture 7: Edge Detection

Lecture 7: Edge Detection #1 Lecture 7: Edge Detection Saad J Bedros sbedros@umn.edu Review From Last Lecture Definition of an Edge First Order Derivative Approximation as Edge Detector #2 This Lecture Examples of Edge Detection

More information

Basics on 2-D 2 D Random Signal

Basics on 2-D 2 D Random Signal Basics on -D D Random Signal Spring 06 Instructor: K. J. Ray Liu ECE Department, Univ. of Maryland, College Park Overview Last Time: Fourier Analysis for -D signals Image enhancement via spatial filtering

More information

Inverse Problems in Image Processing

Inverse Problems in Image Processing H D Inverse Problems in Image Processing Ramesh Neelamani (Neelsh) Committee: Profs. R. Baraniuk, R. Nowak, M. Orchard, S. Cox June 2003 Inverse Problems Data estimation from inadequate/noisy observations

More information

ECE472/572 - Lecture 11. Roadmap. Roadmap. Image Compression Fundamentals and Lossless Compression Techniques 11/03/11.

ECE472/572 - Lecture 11. Roadmap. Roadmap. Image Compression Fundamentals and Lossless Compression Techniques 11/03/11. ECE47/57 - Lecture Image Compression Fundamentals and Lossless Compression Techniques /03/ Roadmap Preprocessing low level Image Enhancement Image Restoration Image Segmentation Image Acquisition Image

More information

A Fourier Transform Model in Excel #1

A Fourier Transform Model in Excel #1 A Fourier Transorm Model in Ecel # -This is a tutorial about the implementation o a Fourier transorm in Ecel. This irst part goes over adjustments in the general Fourier transorm ormula to be applicable

More information

Computer Vision Lecture 3

Computer Vision Lecture 3 Computer Vision Lecture 3 Linear Filters 03.11.2015 Bastian Leibe RWTH Aachen http://www.vision.rwth-aachen.de leibe@vision.rwth-aachen.de Demo Haribo Classification Code available on the class website...

More information

A Comparative Study of Non-separable Wavelet and Tensor-product. Wavelet; Image Compression

A Comparative Study of Non-separable Wavelet and Tensor-product. Wavelet; Image Compression Copyright c 007 Tech Science Press CMES, vol., no., pp.91-96, 007 A Comparative Study o Non-separable Wavelet and Tensor-product Wavelet in Image Compression Jun Zhang 1 Abstract: The most commonly used

More information

Chapter 8 The Discrete Fourier Transform

Chapter 8 The Discrete Fourier Transform Chapter 8 The Discrete Fourier Transform Introduction Representation of periodic sequences: the discrete Fourier series Properties of the DFS The Fourier transform of periodic signals Sampling the Fourier

More information

Edge Detection. Introduction to Computer Vision. Useful Mathematics Funcs. The bad news

Edge Detection. Introduction to Computer Vision. Useful Mathematics Funcs. The bad news Edge Detection Introduction to Computer Vision CS / ECE 8B Thursday, April, 004 Edge detection (HO #5) Edge detection is a local area operator that seeks to find significant, meaningful changes in image

More information

Screen-space processing Further Graphics

Screen-space processing Further Graphics Screen-space processing Rafał Mantiuk Computer Laboratory, University of Cambridge Cornell Box and tone-mapping Rendering Photograph 2 Real-world scenes are more challenging } The match could not be achieved

More information

ENT 315 Medical Signal Processing CHAPTER 2 DISCRETE FOURIER TRANSFORM. Dr. Lim Chee Chin

ENT 315 Medical Signal Processing CHAPTER 2 DISCRETE FOURIER TRANSFORM. Dr. Lim Chee Chin ENT 315 Medical Signal Processing CHAPTER 2 DISCRETE FOURIER TRANSFORM Dr. Lim Chee Chin Outline Introduction Discrete Fourier Series Properties of Discrete Fourier Series Time domain aliasing due to frequency

More information

Application of Wavelet Transform Modulus Maxima in Raman Distributed Temperature Sensors

Application of Wavelet Transform Modulus Maxima in Raman Distributed Temperature Sensors PHOTONIC SENSORS / Vol. 4, No. 2, 2014: 142 146 Application o Wavelet Transorm Modulus Maxima in Raman Distributed Temperature Sensors Zongliang WANG, Jun CHANG *, Sasa ZHANG, Sha LUO, Cuanwu JIA, Boning

More information

EEO 401 Digital Signal Processing Prof. Mark Fowler

EEO 401 Digital Signal Processing Prof. Mark Fowler EEO 401 Digital Signal Processing Pro. Mark Fowler Note Set #10 Fourier Analysis or DT Signals eading Assignment: Sect. 4.2 & 4.4 o Proakis & Manolakis Much o Ch. 4 should be review so you are expected

More information

! Review: Discrete Fourier Transform (DFT) ! DFT Properties. " Duality. " Circular Shift. ! Circular Convolution. ! Fast Convolution Methods

! Review: Discrete Fourier Transform (DFT) ! DFT Properties.  Duality.  Circular Shift. ! Circular Convolution. ! Fast Convolution Methods Toda ESE 531: Digital Signal Processing! Review: Discrete Fourier Transform (DFT)! DFT Properties Lec 20: April 11, 2017 Discrete Fourier Transform, Pt 2 " Dualit " Circular Shift! Circular Convolution!

More information

1. Calculation of the DFT

1. Calculation of the DFT ELE E4810: Digital Signal Processing Topic 10: The Fast Fourier Transform 1. Calculation of the DFT. The Fast Fourier Transform algorithm 3. Short-Time Fourier Transform 1 1. Calculation of the DFT! Filter

More information

Lecture 04 Image Filtering

Lecture 04 Image Filtering Institute of Informatics Institute of Neuroinformatics Lecture 04 Image Filtering Davide Scaramuzza 1 Lab Exercise 2 - Today afternoon Room ETH HG E 1.1 from 13:15 to 15:00 Work description: your first

More information

Digital Image Processing

Digital Image Processing Digital Image Processing 16 November 2006 Dr. ir. Aleksandra Pizurica Prof. Dr. Ir. Wilfried Philips Aleksandra.Pizurica @telin.ugent.be Tel: 09/264.3415 UNIVERSITEIT GENT Telecommunicatie en Informatieverwerking

More information

Contents. Signals as functions (1D, 2D)

Contents. Signals as functions (1D, 2D) Fourier Transform The idea A signal can be interpreted as en electromagnetic wave. This consists of lights of different color, or frequency, that can be split apart usign an optic prism. Each component

More information

! Introduction. ! Discrete Time Signals & Systems. ! Z-Transform. ! Inverse Z-Transform. ! Sampling of Continuous Time Signals

! Introduction. ! Discrete Time Signals & Systems. ! Z-Transform. ! Inverse Z-Transform. ! Sampling of Continuous Time Signals ESE 531: Digital Signal Processing Lec 25: April 24, 2018 Review Course Content! Introduction! Discrete Time Signals & Systems! Discrete Time Fourier Transform! Z-Transform! Inverse Z-Transform! Sampling

More information

Chapter 3: Image Enhancement in the. Office room : 841

Chapter 3: Image Enhancement in the.   Office room : 841 Chapter 3: Image Enhancement in the Spatial Domain Lecturer: Jianbing Shen Email : shenjianbing@bit.edu.cn Oice room : 841 http://cs.bit.edu.cn/shenjianbing cn/shenjianbing Principle Objective o Enhancement

More information

Edge Detection. Image Processing - Computer Vision

Edge Detection. Image Processing - Computer Vision Image Processing - Lesson 10 Edge Detection Image Processing - Computer Vision Low Level Edge detection masks Gradient Detectors Compass Detectors Second Derivative - Laplace detectors Edge Linking Image

More information

Convolution Spatial Aliasing Frequency domain filtering fundamentals Applications Image smoothing Image sharpening

Convolution Spatial Aliasing Frequency domain filtering fundamentals Applications Image smoothing Image sharpening Frequency Domain Filtering Correspondence between Spatial and Frequency Filtering Fourier Transform Brief Introduction Sampling Theory 2 D Discrete Fourier Transform Convolution Spatial Aliasing Frequency

More information

WORCESTER POLYTECHNIC INSTITUTE

WORCESTER POLYTECHNIC INSTITUTE WORCESTER POLYTECHNIC INSTITUTE MECHANICAL ENGINEERING DEPARTMENT Optical Metrology and NDT ME-593L, C 2018 Introduction: Fringe Skeletonization February 2018 Quantitative analysis Fringe skeletonization

More information

IMPROVED NOISE CANCELLATION IN DISCRETE COSINE TRANSFORM DOMAIN USING ADAPTIVE BLOCK LMS FILTER

IMPROVED NOISE CANCELLATION IN DISCRETE COSINE TRANSFORM DOMAIN USING ADAPTIVE BLOCK LMS FILTER SANJAY KUMAR GUPTA* et al. ISSN: 50 3676 [IJESAT] INTERNATIONAL JOURNAL OF ENGINEERING SCIENCE & ADVANCED TECHNOLOGY Volume-, Issue-3, 498 50 IMPROVED NOISE CANCELLATION IN DISCRETE COSINE TRANSFORM DOMAIN

More information

EE123 Digital Signal Processing

EE123 Digital Signal Processing EE123 Digital Signal Processing Lecture 1 Time-Dependent FT Announcements! Midterm: 2/22/216 Open everything... but cheat sheet recommended instead 1am-12pm How s the lab going? Frequency Analysis with

More information

Contents. Signals as functions (1D, 2D)

Contents. Signals as functions (1D, 2D) Fourier Transform The idea A signal can be interpreted as en electromagnetic wave. This consists of lights of different color, or frequency, that can be split apart usign an optic prism. Each component

More information

18/10/2017. Image Enhancement in the Spatial Domain: Gray-level transforms. Image Enhancement in the Spatial Domain: Gray-level transforms

18/10/2017. Image Enhancement in the Spatial Domain: Gray-level transforms. Image Enhancement in the Spatial Domain: Gray-level transforms Gray-level transforms Gray-level transforms Generic, possibly nonlinear, pointwise operator (intensity mapping, gray-level transformation): Basic gray-level transformations: Negative: s L 1 r Generic log:

More information

Multimedia Databases. Previous Lecture. 4.1 Multiresolution Analysis. 4 Shape-based Features. 4.1 Multiresolution Analysis

Multimedia Databases. Previous Lecture. 4.1 Multiresolution Analysis. 4 Shape-based Features. 4.1 Multiresolution Analysis Previous Lecture Multimedia Databases Texture-Based Image Retrieval Low Level Features Tamura Measure, Random Field Model High-Level Features Fourier-Transform, Wavelets Wolf-Tilo Balke Silviu Homoceanu

More information

Multimedia Databases. Wolf-Tilo Balke Philipp Wille Institut für Informationssysteme Technische Universität Braunschweig

Multimedia Databases. Wolf-Tilo Balke Philipp Wille Institut für Informationssysteme Technische Universität Braunschweig Multimedia Databases Wolf-Tilo Balke Philipp Wille Institut für Informationssysteme Technische Universität Braunschweig http://www.ifis.cs.tu-bs.de 4 Previous Lecture Texture-Based Image Retrieval Low

More information

Vision & Perception. Simple model: simple reflectance/illumination model. image: x(n 1,n 2 )=i(n 1,n 2 )r(n 1,n 2 ) 0 < r(n 1,n 2 ) < 1

Vision & Perception. Simple model: simple reflectance/illumination model. image: x(n 1,n 2 )=i(n 1,n 2 )r(n 1,n 2 ) 0 < r(n 1,n 2 ) < 1 Simple model: simple reflectance/illumination model Eye illumination source i(n 1,n 2 ) image: x(n 1,n 2 )=i(n 1,n 2 )r(n 1,n 2 ) reflectance term r(n 1,n 2 ) where 0 < i(n 1,n 2 ) < 0 < r(n 1,n 2 )

More information

Prof. Mohd Zaid Abdullah Room No:

Prof. Mohd Zaid Abdullah Room No: EEE 52/4 Advnced Digital Signal and Image Processing Tuesday, 00-300 hrs, Data Com. Lab. Friday, 0800-000 hrs, Data Com. Lab Prof. Mohd Zaid Abdullah Room No: 5 Email: mza@usm.my www.eng.usm.my Electromagnetic

More information

NEW PIXEL SORTING METHOD FOR PALETTE BASED STEGANOGRAPHY AND COLOR MODEL SELECTION

NEW PIXEL SORTING METHOD FOR PALETTE BASED STEGANOGRAPHY AND COLOR MODEL SELECTION NEW PIXEL SORTING METHOD FOR PALETTE BASED STEGANOGRAPHY AND COLOR MODEL SELECTION Sos S. Agaian 1 and Juan P. Perez 2 1NSP Lab, The University of Texas at San Antonio, Electrical Engineering Department,

More information

Digital Image Processing. Filtering in the Frequency Domain

Digital Image Processing. Filtering in the Frequency Domain 2D Linear Systems 2D Fourier Transform and its Properties The Basics of Filtering in Frequency Domain Image Smoothing Image Sharpening Selective Filtering Implementation Tips 1 General Definition: System

More information

Presenters. Time Domain and Statistical Model Development, Simulation and Correlation Methods for High Speed SerDes

Presenters. Time Domain and Statistical Model Development, Simulation and Correlation Methods for High Speed SerDes JANUARY 28-31, 2013 SANTA CLARA CONVENTION CENTER Time Domain and Statistical Model Development, Simulation and Correlation Methods or High Speed SerDes Presenters Xingdong Dai Fangyi Rao Shiva Prasad

More information

Contents. Signals as functions (1D, 2D)

Contents. Signals as functions (1D, 2D) Fourier Transform The idea A signal can be interpreted as en electromagnetic wave. This consists of lights of different color, or frequency, that can be split apart usign an optic prism. Each component

More information

Fourier analysis of discrete-time signals. (Lathi Chapt. 10 and these slides)

Fourier analysis of discrete-time signals. (Lathi Chapt. 10 and these slides) Fourier analysis of discrete-time signals (Lathi Chapt. 10 and these slides) Towards the discrete-time Fourier transform How we will get there? Periodic discrete-time signal representation by Discrete-time

More information

Chirp Transform for FFT

Chirp Transform for FFT Chirp Transform for FFT Since the FFT is an implementation of the DFT, it provides a frequency resolution of 2π/N, where N is the length of the input sequence. If this resolution is not sufficient in a

More information

Fall 2011, EE123 Digital Signal Processing

Fall 2011, EE123 Digital Signal Processing Lecture 6 Miki Lustig, UCB September 11, 2012 Miki Lustig, UCB DFT and Sampling the DTFT X (e jω ) = e j4ω sin2 (5ω/2) sin 2 (ω/2) 5 x[n] 25 X(e jω ) 4 20 3 15 2 1 0 10 5 1 0 5 10 15 n 0 0 2 4 6 ω 5 reconstructed

More information

TLT-5200/5206 COMMUNICATION THEORY, Exercise 3, Fall TLT-5200/5206 COMMUNICATION THEORY, Exercise 3, Fall Problem 1.

TLT-5200/5206 COMMUNICATION THEORY, Exercise 3, Fall TLT-5200/5206 COMMUNICATION THEORY, Exercise 3, Fall Problem 1. TLT-5/56 COMMUNICATION THEORY, Exercise 3, Fall Problem. The "random walk" was modelled as a random sequence [ n] where W[i] are binary i.i.d. random variables with P[W[i] = s] = p (orward step with probability

More information

VII. Discrete Fourier Transform (DFT) Chapter-8. A. Modulo Arithmetic. (n) N is n modulo N, n is an integer variable.

VII. Discrete Fourier Transform (DFT) Chapter-8. A. Modulo Arithmetic. (n) N is n modulo N, n is an integer variable. 1 VII. Discrete Fourier Transform (DFT) Chapter-8 A. Modulo Arithmetic (n) N is n modulo N, n is an integer variable. (n) N = n m N 0 n m N N-1, pick m Ex. (k) 4 W N = e -j2π/n 2 Note that W N k = 0 but

More information

Local Enhancement. Local enhancement

Local Enhancement. Local enhancement Local Enhancement Local Enhancement Median filtering (see notes/slides, 3.5.2) HW4 due next Wednesday Required Reading: Sections 3.3, 3.4, 3.5, 3.6, 3.7 Local Enhancement 1 Local enhancement Sometimes

More information

2D Discrete Fourier Transform (DFT)

2D Discrete Fourier Transform (DFT) 2D Discrete Fourier Transform (DFT) Outline Circular and linear convolutions 2D DFT 2D DCT Properties Other formulations Examples 2 2D Discrete Fourier Transform Fourier transform of a 2D signal defined

More information

Midterm Review. Yao Wang Polytechnic University, Brooklyn, NY 11201

Midterm Review. Yao Wang Polytechnic University, Brooklyn, NY 11201 Midterm Review Yao Wang Polytechnic University, Brooklyn, NY 11201 Based on J. L. Prince and J. M. Links, Medical maging Signals and Systems, and lecture notes by Prince. Figures are from the textbook.

More information

Design of Coprime DFT Arrays and Filter Banks

Design of Coprime DFT Arrays and Filter Banks Design o Coprime DFT Arrays and Filter Banks Chun-Lin Liu clliu@caltechedu P P Vaidyanathan ppvnath@systemscaltechedu Digital Signal Processing Group Electrical Engineering Caliornia Institute o Technology

More information

Lecture 20: Discrete Fourier Transform and FFT

Lecture 20: Discrete Fourier Transform and FFT EE518 Digital Signal Processing University of Washington Autumn 2001 Dept of Electrical Engineering Lecture 20: Discrete Fourier Transform and FFT Dec 10, 2001 Prof: J Bilmes TA:

More information

Ex x xf xdx. Ex+ a = x+ a f x dx= xf x dx+ a f xdx= xˆ. E H x H x H x f x dx ˆ ( ) ( ) ( ) μ is actually the first moment of the random ( )

Ex x xf xdx. Ex+ a = x+ a f x dx= xf x dx+ a f xdx= xˆ. E H x H x H x f x dx ˆ ( ) ( ) ( ) μ is actually the first moment of the random ( ) Fall 03 Analysis o Eperimental Measurements B Eisenstein/rev S Errede The Epectation Value o a Random Variable: The epectation value E[ ] o a random variable is the mean value o, ie ˆ (aa μ ) For discrete

More information

Least-Squares Spectral Analysis Theory Summary

Least-Squares Spectral Analysis Theory Summary Least-Squares Spectral Analysis Theory Summary Reerence: Mtamakaya, J. D. (2012). Assessment o Atmospheric Pressure Loading on the International GNSS REPRO1 Solutions Periodic Signatures. Ph.D. dissertation,

More information

Intensity Transformations and Spatial Filtering: WHICH ONE LOOKS BETTER? Intensity Transformations and Spatial Filtering: WHICH ONE LOOKS BETTER?

Intensity Transformations and Spatial Filtering: WHICH ONE LOOKS BETTER? Intensity Transformations and Spatial Filtering: WHICH ONE LOOKS BETTER? : WHICH ONE LOOKS BETTER? 3.1 : WHICH ONE LOOKS BETTER? 3.2 1 Goal: Image enhancement seeks to improve the visual appearance of an image, or convert it to a form suited for analysis by a human or a machine.

More information

EECS490: Digital Image Processing. Lecture #11

EECS490: Digital Image Processing. Lecture #11 Lecture #11 Filtering Applications: OCR, scanning Highpass filters Laplacian in the frequency domain Image enhancement using highpass filters Homomorphic filters Bandreject/bandpass/notch filters Correlation

More information

Chapter 4 Imaging. Lecture 21. d (110) Chem 793, Fall 2011, L. Ma

Chapter 4 Imaging. Lecture 21. d (110) Chem 793, Fall 2011, L. Ma Chapter 4 Imaging Lecture 21 d (110) Imaging Imaging in the TEM Diraction Contrast in TEM Image HRTEM (High Resolution Transmission Electron Microscopy) Imaging or phase contrast imaging STEM imaging a

More information

Signals & Linear Systems Analysis Chapter 2&3, Part II

Signals & Linear Systems Analysis Chapter 2&3, Part II Signals & Linear Systems Analysis Chapter &3, Part II Dr. Yun Q. Shi Dept o Electrical & Computer Engr. New Jersey Institute o echnology Email: shi@njit.edu et used or the course:

More information

Chapter 16. Local Operations

Chapter 16. Local Operations Chapter 16 Local Operations g[x, y] =O{f[x ± x, y ± y]} In many common image processing operations, the output pixel is a weighted combination of the gray values of pixels in the neighborhood of the input

More information

Multimedia Databases. 4 Shape-based Features. 4.1 Multiresolution Analysis. 4.1 Multiresolution Analysis. 4.1 Multiresolution Analysis

Multimedia Databases. 4 Shape-based Features. 4.1 Multiresolution Analysis. 4.1 Multiresolution Analysis. 4.1 Multiresolution Analysis 4 Shape-based Features Multimedia Databases Wolf-Tilo Balke Silviu Homoceanu Institut für Informationssysteme Technische Universität Braunschweig http://www.ifis.cs.tu-bs.de 4 Multiresolution Analysis

More information