October 31, 2006Thesis Defense, UTK1/30 Variational and Partial Differential Equation Models for Color Image Denoising and Their Numerical Approximation.

Slides:



Advertisements
Similar presentations
Finite Difference Discretization of Hyperbolic Equations: Linear Problems Lectures 8, 9 and 10.
Advertisements

Total Variation and Geometric Regularization for Inverse Problems
Johann Radon Institute for Computational and Applied Mathematics: 1/25 Signal- und Bildverarbeitung, Image Analysis and Processing.
Image Enhancement by Regularization Methods Andrey S. Krylov, Andrey V. Nasonov, Alexey S. Lukin Moscow State University Faculty of Computational Mathematics.
Various Regularization Methods in Computer Vision Min-Gyu Park Computer Vision Lab. School of Information and Communications GIST.
Edge Preserving Image Restoration using L1 norm
TVL1 Models for Imaging: Global Optimization & Geometric Properties Part I Tony F. Chan Math Dept, UCLA S. Esedoglu Math Dept, Univ. Michigan Other Collaborators:
Bregman Iterative Algorithms for L1 Minimization with
Topic 6 - Image Filtering - I DIGITAL IMAGE PROCESSING Course 3624 Department of Physics and Astronomy Professor Bob Warwick.
2013 SIAM Great Lakes Section From PDEs to Information Science and Back Russel Caflisch IPAM Mathematics Department, UCLA 1.
P. Venkataraman Mechanical Engineering P. Venkataraman Rochester Institute of Technology DETC2013 – 12269: Continuous Solution for Boundary Value Problems.
Martin Burger Total Variation 1 Cetraro, September 2008 Variational Methods and their Analysis Questions: - Existence - Uniqueness - Optimality conditions.
P. Venkataraman Mechanical Engineering P. Venkataraman Rochester Institute of Technology DETC2014 – 35148: Continuous Solution for Boundary Value Problems.
Tracking Unknown Dynamics - Combined State and Parameter Estimation Tracking Unknown Dynamics - Combined State and Parameter Estimation Presenters: Hongwei.
P. Venkataraman Mechanical Engineering P. Venkataraman Rochester Institute of Technology DETC2011 –47658 Determining ODE from Noisy Data 31 th CIE, Washington.
Yves Meyer’s models for image decomposition and computational approaches Luminita Vese Department of Mathematics, UCLA Triet Le (Yale University), Linh.
ECE 472/572 - Digital Image Processing Lecture 8 - Image Restoration – Linear, Position-Invariant Degradations 10/10/11.
Martin Burger Institut für Numerische und Angewandte Mathematik CeNoS Level set methods for imaging and application to MRI segmentation.
MA5233: Computational Mathematics
Finite Element Method Introduction General Principle
1 Total variation minimization Numerical Analysis, Error Estimation, and Extensions Martin Burger Johannes Kepler University Linz SFB Numerical-Symbolic-Geometric.
Real-time Combined 2D+3D Active Appearance Models Jing Xiao, Simon Baker,Iain Matthew, and Takeo Kanade CVPR 2004 Presented by Pat Chan 23/11/2004.
Optical Flow Brightness Constancy The Aperture problem Regularization Lucas-Kanade Coarse-to-fine Parametric motion models Direct depth SSD tracking.
Total Variation Imaging followed by spectral decomposition using continuous wavelet transform Partha Routh 1 and Satish Sinha 2, 1 Boise State University,
Error Estimation in TV Imaging Martin Burger Institute for Computational and Applied Mathematics European Institute for Molecular Imaging (EIMI) Center.
Weak Formulation ( variational formulation)
SUSAN: structure-preserving noise reduction EE264: Image Processing Final Presentation by Luke Johnson 6/7/2007.
Optical flow and Tracking CISC 649/849 Spring 2009 University of Delaware.
Optical Flow Brightness Constancy The Aperture problem Regularization Lucas-Kanade Coarse-to-fine Parametric motion models Direct depth SSD tracking.
1 Bayesian Restoration Using a New Nonstationary Edge-Preserving Image Prior Giannis K. Chantas, Nikolaos P. Galatsanos, and Aristidis C. Likas IEEE Transactions.
Optical Flow Estimation using Variational Techniques Darya Frolova.
3D Rigid/Nonrigid RegistrationRegistration 1)Known features, correspondences, transformation model – feature basedfeature based 2)Specific motion type,
EE565 Advanced Image Processing Copyright Xin Li Different Frameworks for Image Processing Statistical/Stochastic Models: Wiener’s MMSE estimation.
Mehdi Ghayoumi MSB rm 132 Ofc hr: Thur, a Machine Learning.
6 6.3 © 2012 Pearson Education, Inc. Orthogonality and Least Squares ORTHOGONAL PROJECTIONS.
Computer Vision Optical Flow Marc Pollefeys COMP 256 Some slides and illustrations from L. Van Gool, T. Darell, B. Horn, Y. Weiss, P. Anandan, M. Black,
Types of Governing equations
Numerical methods for PDEs PDEs are mathematical models for –Physical Phenomena Heat transfer Wave motion.
Linear Algebra and Image Processing
Unitary Extension Principle: Ten Years After Zuowei Shen Department of Mathematics National University of Singapore.
Motivation from Real-World Applications EE565 Advanced Image Processing Copyright Xin Li Noisy Photos Noisy ultrasound data.
IIS for Image Processing Michael J. Watts
1/20 Obtaining Shape from Scanning Electron Microscope Using Hopfield Neural Network Yuji Iwahori 1, Haruki Kawanaka 1, Shinji Fukui 2 and Kenji Funahashi.
PDE-based Methods for Image and Shape Processing Applications Alexander Belyaev School of Engineering & Physical Sciences Heriot-Watt University, Edinburgh.
7.1. Mean Shift Segmentation Idea of mean shift:
Shape from Shading and Texture. Lambertian Reflectance Model Diffuse surfaces appear equally bright from all directionsDiffuse surfaces appear equally.
Total Variation and Euler's Elastica for Supervised Learning
CHAPTER 5 S TOCHASTIC G RADIENT F ORM OF S TOCHASTIC A PROXIMATION Organization of chapter in ISSO –Stochastic gradient Core algorithm Basic principles.
Effective Optical Flow Estimation
Over-Parameterized Variational Optical Flow
1 Markov random field: A brief introduction (2) Tzu-Cheng Jen Institute of Electronics, NCTU
Partial Derivatives Example: Find If solution: Partial Derivatives Example: Find If solution: gradient grad(u) = gradient.
Head Segmentation using a finite element approach J.Piovano, T. Papadopoulo Odyssée Laboratory (ENPC, ENS, INRIA), INRIA, Sophia-Antipolis, France I. Introduction.
Bivariate Splines for Image Denoising*° *Grant Fiddyment University of Georgia, 2008 °Ming-Jun Lai Dept. of Mathematics University of Georgia.
Air Systems Division Definition of anisotropic denoising operators via sectional curvature Stanley Durrleman September 19, 2006.
CS654: Digital Image Analysis Lecture 22: Image Restoration.
Finite Element Method. History Application Consider the two point boundary value problem.
Using Neumann Series to Solve Inverse Problems in Imaging Christopher Kumar Anand.
Amir Yavariabdi Introduction to the Calculus of Variations and Optical Flow.
Introduction to Medical Imaging Week 6: Introduction to Medical Imaging Week 6: Denoising (part II) – Variational Methods and Evolutions Guy Gilboa Course.
Proposed Courses. Important Notes State-of-the-art challenges in TV Broadcasting o New technologies in TV o Multi-view broadcasting o HDR imaging.
L ECTURE 3 PDE Methods for Image Restoration. O VERVIEW Generic 2 nd order nonlinear evolution PDE Classification: Forward parabolic (smoothing): heat.
1. 2 What is Digital Image Processing? The term image refers to a two-dimensional light intensity function f(x,y), where x and y denote spatial(plane)
MAN-522 Computer Vision Spring
PDE Methods for Image Restoration
Peter Benner and Lihong Feng
Image Restoration and Denoising
Finite Element Surface-Based Stereo 3D Reconstruction
Advanced deconvolution techniques and medical radiography
Shape from Shading and Texture
Presentation transcript:

October 31, 2006Thesis Defense, UTK1/30 Variational and Partial Differential Equation Models for Color Image Denoising and Their Numerical Approximation using Finite Element Methods Thesis Defense Miun Yoon

October 31, 2006Thesis Defense, UTK2/30 Digital Image Processing - Image restoration - Image compression -Image segmentation What is “Digital Image Processing”? observed image Stochastic modeling Wavelets Variational & PDE modeling output

October 31, 2006Thesis Defense, UTK3/30 What is an image in Mathematics? Pixel: Picture + Element observed image color image

October 31, 2006Thesis Defense, UTK4/30 Pixel Representations RGB Color Image :256 shades of RGB Gray Image :256 shades of gray- level

October 31, 2006Thesis Defense, UTK5/30 Image Denoising Model original image “unknown” additive noise noisy image

October 31, 2006Thesis Defense, UTK6/30 Gray Image Denoising Total Variational (TV) Model: Rudin, Osher, and Fatemi [Rud92](1992) Constrained minimization problem: Constraints noisy image Error level

October 31, 2006Thesis Defense, UTK7/30 Unconstrained minimization problem: Gradient Flow (TV Flow):

October 31, 2006Thesis Defense, UTK8/30 Regularized Problem Previous Studies - A. Chamnbolle and P. –L. Lions [Cha97]: proved the existence and the uniqueness result for constraint minimization problem and unconstraint minimization problem is equivalent to the constraint minimization for a unique and non-negative - X. Feng and A. Prohl [Fen03]: proved the existence and the uniqueness for the TV flow and regularized problem and an error analysis for the fully discrete finite approximation for the regularized problem

October 31, 2006Thesis Defense, UTK9/30 Weak Formulation

October 31, 2006Thesis Defense, UTK10/30 Semi-Discrete Finite Element Method T h = {K 1,…,K mR } Finite-Dimensional subspace : set of all vertices of the triangulation T h uniquely determined & forms a basis for V h

October 31, 2006Thesis Defense, UTK11/30 Semi-Discrete Finite Element Method Non-linear ODE system in  t 

October 31, 2006Thesis Defense, UTK12/30 Fully Discrete Finite Element Method X. Feng, M. von Oehen and A. Prohl [Fen05]: rate of convergence for the fully discrete finite approximation of the regularized problem

October 31, 2006Thesis Defense, UTK13/30 Numerical Tests I t=0 t=5e-5 t=1e-4t=1.5e-4 t=2e-4

October 31, 2006Thesis Defense, UTK14/30 Numerical Tests II t=0 t=5e-5 t=1e-4t=1.5e-4 t=2e-4

October 31, 2006Thesis Defense, UTK15/30 Numerical Tests III t=0 t=5e-5 t=1e-4t=1.5e-4 t=2e-4

October 31, 2006Thesis Defense, UTK16/30 Color Image Denoising brightness chromaticity color vector TV flow P-harmonic map flow Non-flat feature channel by channel model chromaticity & brightness (CB) Model

October 31, 2006Thesis Defense, UTK17/30 p-harmonic Map Minimizer of E p Euler-Lagrange equation unit sphere p-energy Constrained Minimization Problem constraint p-harmonic map

October 31, 2006Thesis Defense, UTK18/30 p-harmonic Color Image Denoising Model Gradient flow Non-linear Non-convex Non-linear

October 31, 2006Thesis Defense, UTK19/30 Regularization of p-energy nonlinearnonconvex

October 31, 2006Thesis Defense, UTK20/30 Regularized Model p-harmonic map heat flow

October 31, 2006Thesis Defense, UTK21/30 Weak Formulation

October 31, 2006Thesis Defense, UTK22/30 Semi-Discrete Finite Element Method : set of all vertices of the triangulation T h Finite-Dimensional subspace T h = {K 1,…,K mR }

October 31, 2006Thesis Defense, UTK23/30 Semi-Discrete Finite Element Method Non-linear ODE system in  t 

October 31, 2006Thesis Defense, UTK24/30 Semi-Discrete Finite Element Method

October 31, 2006Thesis Defense, UTK25/30 Fully-Discrete Finite Element Method Decomposition of the density function

October 31, 2006Thesis Defense, UTK26/30 Numerical Tests t=0t=2e-4 t=5e-4t=7e-4t=1e-3

October 31, 2006Thesis Defense, UTK27/30 Generalization Generalized model of the p-harmonic map Regularized flow of generalized model

October 31, 2006Thesis Defense, UTK28/30 Numerical Tests I and q=1 t=2e-4 t=5e-4 t=7e-4 t=1e-3

October 31, 2006Thesis Defense, UTK29/30 Numerical Tests II and q=1 t=2e-4 t=5e-4 t=7e-4 t=1e-3

October 31, 2006Thesis Defense, UTK30/30 Numerical Tests III channel-by-channel t=1e-4 t=3e-4 t=5e-4 t=1e-3

31 Appendix