Scribd
Upload
27 views6 pages

CP2 NP2023

1) The document describes solving the Perona-Malik equation for nonlinear diffusion using numerical methods to process digital images. 2) Two discretization methods must be chosen from a list to solve the diffusion equations: Adams-Moulton, Adams-Bashforth, explicit Runge-Kutta or diagonally implicit Runge-Kutta. 3) The methods are used to reduce noise in a digital image by applying the Perona-Malik diffusion equation.

Uploaded by

giorgiclash282
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
27 views6 pages

CP2 NP2023

1) The document describes solving the Perona-Malik equation for nonlinear diffusion using numerical methods to process digital images. 2) Two discretization methods must be chosen from a list to solve the diffusion equations: Adams-Moulton, Adams-Bashforth, explicit Runge-Kutta or diagonally implicit Runge-Kutta. 3) The methods are used to reduce noise in a digital image by applying the Perona-Malik diffusion equation.

Uploaded by

giorgiclash282
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

Numerical Programming

Ramaz Botchorishvili

Kutaisi International University

December 21, 2023

Ramaz Botchorishvili (KIU) Computational Project #1 Fall Term 2023 1/6


Nonlinear Diffusion and Digital Images
Problem 2.1
Solve Nonlinear diffusion and apply to digital images
1. Consider method of lines for Perona-Malik equation
2. Discretize obtained ODEs using two different methods from the list
below:
2.1 Adams-Moulton method
2.2 Adams-Bashforth method
2.3 Explicit Runge-Kutta method
2.4 Diagonally implicit Runge-Kutta method
3. Truncation error of the selected methods should be greater than two.
Check truncation error of the chosen methods.
4. Check stability of the chosen methods
5. Choose digital image and add noise
6. Remove/reduce noise using Perona-Malik equation and selected
methods.
7. Conclusion: formulate criteria and compare methods. Does parameter
Ramaz Botchorishvili (KIU) Computational Project #1 Fall Term 2023 2/6
Perona-Malik equation and digital images

Important notice
1. Each item/sub-item above is worth of 1 point
2. Total points - 12
3. The problem/sub-problem is assigned 0 point if
▶ same set of student defined parameters are used by two or more students
▶ results cannot be replicated
▶ your code fails: does not produce correct results on new tests given by
instructor
▶ your code does not compile
▶ good documentation of your code is missing

Ramaz Botchorishvili (KIU) Computational Project #1 Fall Term 2023 3/6


Perona-Malik Equation

▶ Nonlinear diffusion model for image processing


▶ Preserves edges while smoothing homogeneous regions
▶ Equation:
∂u
= div (g(∥∇u∥)∇u)
∂t
▶ u(x , y , t): image intensity
▶ g(s): diffusion coefficient

Ramaz Botchorishvili (KIU) Computational Project #1 Fall Term 2023 4/6


Diffusion Coefficient g(s)

▶ Controls edge preservation


▶ Common choices:
▶ g(s) = 1/(1 + s 2 /k 2 ) (Perona-Malik)
▶ g(s) = exp(−s 2 /k 2 ) (Charbonnier)
 2
− 21
▶ g(s) = 1 + ks 2 (Weickert)
1
▶ g(s) = 1+(s/k)α (Generalized Perona-Malik)
▶ g(s) = 1−exp(−s 2 /k 2 )
s 2 /k 2 (Yule-Nielsen)
▶ Parameter k regulates edge sensitivity
▶ Different choices lead to varying edge preservation and noise reduction
characteristics

Ramaz Botchorishvili (KIU) Computational Project #1 Fall Term 2023 5/6


Numerical Schemes

▶ Forward Euler discretization, example for 2D case:


n n n ) − gn
− ui,j n n
n+1 n
gi+1/2,j (ui+1,j i−1/2,j (ui,j − ui−1,j )
ui,j = ui,j + ∆t
h2
n n n ) − gn
− ui,j n n
i,j−1/2 (ui,j − ui,j−1 )
!
gi,j+1/2 (ui,j+1
+
h2

▶ Method of lines, semi-discretization, example for 2D case:

dui,j (t) gi+1/2,j (ui+1,j (t) − ui,j (t)) − gi−1/2,j (ui,j (t) − ui−1,j (t))
=
dt h2
gi,j+1/2 (ui,j+1 (t) − ui,j (t)) − gi,j−1/2 (ui,j (t) − ui,j−1 (t))
+
h2

Ramaz Botchorishvili (KIU) Computational Project #1 Fall Term 2023 6/6

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy