Fourier Transforms 1D

Similar documents
Fourier Transform 4: z-transform (part 2) & Introduction to 2D Fourier Analysis

ECE Digital Image Processing and Introduction to Computer Vision. Outline

Introduction to Computer Vision. 2D Linear Systems

Image Enhancement: Methods. Digital Image Processing. No Explicit definition. Spatial Domain: Frequency Domain:

Digital Image Processing COSC 6380/4393

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

Empirical Mean and Variance!

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

E : Lecture 1 Introduction

Computer Vision & Digital Image Processing

ITK Filters. Thresholding Edge Detection Gradients Second Order Derivatives Neighborhood Filters Smoothing Filters Distance Map Image Transforms

Reading. 3. Image processing. Pixel movement. Image processing Y R I G Q

Digital Signal Processing Lecture 10 - Discrete Fourier Transform

Histogram Processing

Image preprocessing in spatial domain

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

Digital Image Processing COSC 6380/4393

Chapter 4: Filtering in the Frequency Domain. Fourier Analysis R. C. Gonzalez & R. E. Woods

Digital Image Processing. Filtering in the Frequency Domain

Sampling in 1D ( ) Continuous time signal f(t) Discrete time signal. f(t) comb

IMAGE ENHANCEMENT II (CONVOLUTION)

Chirp images in 2-D fractional Fourier transform domain. Presenter: Ming-Feng Lu Beijing Institute of Technology July 20, 2016

Local Enhancement. Local enhancement

Local enhancement. Local Enhancement. Local histogram equalized. Histogram equalized. Local Contrast Enhancement. Fig 3.23: Another example

Chapter-2 Relations and Functions. Miscellaneous

Lecture 3: Linear Filters

From Fourier Series to Analysis of Non-stationary Signals - II

Review of Analog Signal Analysis

Lecture 3: Linear Filters

Image Enhancement in the frequency domain. GZ Chapter 4

Spatial Enhancement Region operations: k'(x,y) = F( k(x-m, y-n), k(x,y), k(x+m,y+n) ]

Filtering in the Frequency Domain

ECG782: Multidimensional Digital Signal Processing

CITS 4402 Computer Vision

Linear Operators and Fourier Transform

COMP344 Digital Image Processing Fall 2007 Final Examination

Wavelets and Multiresolution Processing

Machine vision. Summary # 4. The mask for Laplacian is given

Linear Filters and Convolution. Ahmed Ashraf

G52IVG, School of Computer Science, University of Nottingham

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

Each problem is worth 25 points, and you may solve the problems in any order.

Computer Vision. Filtering in the Frequency Domain

Machine vision, spring 2018 Summary 4

Tutorial Sheet #2 discrete vs. continuous functions, periodicity, sampling

Today s lecture. Local neighbourhood processing. The convolution. Removing uncorrelated noise from an image The Fourier transform

Chap 4. Sampling of Continuous-Time Signals

Computer Vision Lecture 3

Image Acquisition and Sampling Theory

Images have structure at various scales

Inverse Problems in Image Processing

EE 261 The Fourier Transform and its Applications Fall 2006 Midterm Exam Solutions

The objective of this LabVIEW Mini Project was to understand the following concepts:

Multiscale Image Transforms

Sensors. Chapter Signal Conditioning

DEPARTMENT OF ELECTRICAL AND ELECTRONIC ENGINEERING EXAMINATIONS 2010

Two-Dimensional Signal Processing and Image De-noising

Convolution. Define a mathematical operation on discrete-time signals called convolution, represented by *. Given two discrete-time signals x 1, x 2,

3.2 Complex Sinusoids and Frequency Response of LTI Systems

Filtering in the Frequency Domain

MIT 2.71/2.710 Optics 10/31/05 wk9-a-1. The spatial frequency domain

IB Paper 6: Signal and Data Analysis

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

Filtering in Frequency Domain

ECE Digital Image Processing and Introduction to Computer Vision. Outline

Chapter 16. Local Operations

FROM ANALOGUE TO DIGITAL

Basics on 2-D 2 D Random Signal

Signals, Instruments, and Systems W5. Introduction to Signal Processing Sampling, Reconstruction, and Filters

EXAMINATION QUESTION PAPER

Chapter 1 Fundamental Concepts

Image Processing. Transforms. Mylène Christine Queiroz de Farias

Image Processing. Waleed A. Yousef Faculty of Computers and Information, Helwan University. April 3, 2010

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

DHANALAKSHMI COLLEGE OF ENGINEERING DEPARTMENT OF ELECTRICAL AND ELECTRONICS ENGINEERING EC2314- DIGITAL SIGNAL PROCESSING UNIT I INTRODUCTION PART A

Digital Filters Ying Sun

Image Filtering, Edges and Image Representation

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

Review: Continuous Fourier Transform

Signal Processing COS 323

Module 3 LOSSY IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur

Solutions to Problems in Chapter 4

CAP 5415 Computer Vision Fall 2011

ELEG 3124 SYSTEMS AND SIGNALS Ch. 5 Fourier Transform

Introduction to Image Processing #5/7

ANALOG AND DIGITAL SIGNAL PROCESSING ADSP - Chapter 8

Chapter 8 The Discrete Fourier Transform

CAP 5415 Computer Vision

Fourier Analysis and Imaging Ronald Bracewell L.M. Terman Professor of Electrical Engineering Emeritus Stanford University Stanford, California

PS403 - Digital Signal processing

Principles of Communications

Continuous-Time Fourier Transform

Review of Linear Time-Invariant Network Analysis

Outline. Convolution. Filtering

VU Signal and Image Processing

6.869 Advances in Computer Vision. Bill Freeman, Antonio Torralba and Phillip Isola MIT Oct. 3, 2018

Deterministic sampling masks and compressed sensing: Compensating for partial image loss at the pixel level

6.02 Fall 2012 Lecture #11

The Frequency Domain, without tears. Many slides borrowed from Steve Seitz

Convolution and Linear Systems

Transcription:

Fourier Transforms 1D 3D Image Processing Alireza Ghane 1

Overview Recap Intuitions Function representations shift-invariant spaces linear, time-invariant (LTI) systems complex numbers Fourier Transforms Transform pairs properties Convolution Theorem Understanding sampling in the Fourier space 2

What is a (digital) image? An image is made of pixels (=picture elements) the coordinate values are discretized Laurent Condat / Torsten Möller 3

Quantization The pixel values are quantized: they belong to a discrete set of values, generally represented by integers between 0 and N-1 coded on 8 bits 255 possible val. coded on 4 bits Laurent Condat / Torsten Möller coded on 2 bits 4 possible val. 4

2D lattices An image is defined on a lattice. The most common is the Cartesian (a.k.a square) lattice. But other lattices exist and have interesting properties. Laurent Condat / Torsten Möller 5

Image quality Common defaults in images: blur (motion blur, out of focus blur...) ringing aliasing (staircasing, Moiré patterns) Laurent Condat / Torsten Möller 6

Image quality Low-frequency Moiré artifacts appear when high-frequency content is sampled in an incorrect way. Laurent Condat / Torsten Möller 7

Image quality We need to be able to measure the difference between two images, for instance an original image and a degraded one. Classical difference measures between two images I 1 and I 2 : mean absolute error: MAE = 1 P W P H WH k x =1 k y =1 I 1[k x,k y ] I 2 [k x,k y ] mean square error: MSE = 1 P W P H WH k x =1 k y =1 I 1[k x,k y ] I 2 [k x,k y ] 2 255 2 peak signal to noise ratio (db): PSNR = 10 log 10 MSE There exist much more sophisticated quality measures and difference measures for images (SSIM...) Laurent Condat / Torsten Möller 8

Image enhancement Intensity Transformations image negatives log transforms gamma (power-law) transforms contrast stretching intensity-level slicing bit-plane slicing 9

Histogram equalization Idea -- stretch histogram non-uniformally such that final histogram is a uniform distribution p s (s) =p r (r) dr ds s = T (r) = Z r 0 p r (w)dw 10

Histogram matching original matched equalization 11

Local histogram equalization 12

Mechanics of filtering Correlation: w(x, y)? f(x, y) = ax bx w(s, t) f(x + s, y + t) Convolution: w(x, y) f(x, y) = s= a ax t= b bx w(s, t) f(x s, y t) s= a t= b 13

Median filtering (denoising) 14

Sharpening enhance / highlight transition in intensity how to find transition? unsharp masking / highboost filtering first / second order derivatives in 1D multi-d: gradient magnitude Laplacian 15

Overview Recap Intuitions Function representations shift-invariant spaces linear, time-invariant (LTI) systems complex numbers Fourier Transforms Transform pairs properties Convolution Theorem Understanding sampling in the Fourier space 16

Basis Vectors P P =3i +2j j b a P =1.6a +1.9b 0 i Alireza Ghane 17

Basis Vectors Linear combination of the basis vectors can express any point in the point space [grey rectangle]. P P =3i +2j j b a P =1.6a +1.9b 0 i Alireza Ghane 18

Basis Functions Linear combination of the basis functions can express any function in a the function space (subset of the functions world). Example: v k (t) = 1 t = k 0 otherwise 0 k f(t) = X k2r f(t k )v k (t) Alireza Ghane 19

Time vs. Frequency Space y = sin(2 100t) Y (f) ={ (f == ±100Hz)?1:0} y = sin(2 800t) Y (f) ={ (f == ±800Hz)?1:0} y = sin(2 100t)+sin(2 800t) Y (f) ={ (f == ±100Hz ±800Hz)?1:0} Alireza Ghane 20

Time vs. Frequency Space y = sin(2 100t) Y (f) ={ (f == ±100Hz)?1:0} y = sin(2 1t) Y (f) ={ (f == ±1Hz)?1:0} y = sin(2 1t) sin(1 100t) Y (f) ={ (f == ±99Hz ±101Hz)?1:0} y =0.5[cos(2 99t) cos(2 101t)] Alireza Ghane 21

Building Square Wave + + + + Alireza Ghane 22

Overview Recap Intuitions Function representations shift-invariant spaces linear, time-invariant (LTI) systems complex numbers Fourier Transforms Transform pairs properties Convolution Theorem Understanding sampling in the Fourier space 23

How to represent a function? on a computer, can only store a bunch of numbers, not a continuous function! brute-force idea: sampling 24

What is sampling? (mathematically speaking) modelled through an impulse not really a function, but a distribution: Z 1 1 (t) = 1 if t =0 (t)dt =1 0 if t 6= 0 25

The sifting property picking a value off from f: more general: Z 1 1 Z 1 1 f(t) (t)dt = f(0) f(t) (t t 0 )dt = f(t 0 ) 26

The impulse train pick up multiple values of f at once: s T (t) = 1X 1 (t n T ) 27

What is sampling? f(t) s T (t) = 1X f(n T ) (t n T ) 1 1X 1 f[n] (t n T ) 28