Radial Basis Functions and Application in Edge Detection

1
Radial Basis Functions and Application in Edge Detection Chris Cacciatore, Tian Jiang, and Kerenne Paul University of Massachusetts Dartmouth Department of Mathematics Saeja Kim and Sigal Gottlieb Abstract This project focuses on the use of Radial Basis Functions in Edge Detection in two- dimensional images. We will be using a 2-D iterative RBF edge detection method. We will be varying the point distribution and shape parameter while also quantifying the effects of the accuracy of the edge detection on 2-D images. Furthermore, we study a variety of Radial Basis Functions and their accuracy in Edge Detection. Radial Basis Functions (RBF’s) Radial Basis Functions use the distance between two points on a given interval and epsilon (shape parameter) as variables. Three common types of RBF’s are Multi-quadric, Inverse Multi- quadric, and Gaussian. Multi-quadric •= Gaussian • Exp() This method changes the values of the shape parameters depending on the smoothness of f(x). Using this method allows the accuracy of the approximations to be solely determined on . The Main idea is that disappears only near the center of the discontinuity resulting in the basis functions near the discontinuity to become We have used three different radial basis functions and their results at different values of epsilon. Each function requires different values of epsilon to render a recognizable edge map. Future Work Explore further into matrix involvement in edge detection Look into effects different parts of the code, TwoD_Example1, have on edge maps Test for necessary changes in epsilon for different sized RBF’s and their Results Original image M = zeros(N); MD = M; for ix = 1:N for iy = 1:N M(ix,iy) = sqrt( (x(ix)-x(iy))^2 + (eps(iy))^2); if M(ix,iy) == 0 MD(ix,iy) = 0; else MD(ix,iy) = (x(ix) - x(iy))/M(ix,iy); end Multi-quadric Inverse Multi- quadric M = zeros(N); MD = M; for ix = 1:N for iy = 1:N M(ix,iy) = 1/sqrt( (x(ix)-x(iy))^2 + (eps(iy))^2); if M(ix,iy) == 0 MD(ix,iy) = 0; else MD(ix,iy) = -(x(ix) - x(iy))/sqrt( ((x(ix)-x(iy))^2 + (eps(iy))^2)^3); end Gaussian M = zeros(N); MD = M; for ix = 1:N for iy = 1:N M(ix,iy) = exp(-((eps(iy))^2)*((x(ix)- x(iy))^2)); if M(ix,iy) == 0 MD(ix,iy) = 0; else MD(ix,iy) = -2*((eps(iy))^2)*(x(ix)- x(iy))*exp(-((eps(iy))^2)*(x(ix)- x(iy))^2); end Multi-quadric Gaussian Inverse Multi- quadric Original Image

description

Radial Basis Functions and Application in Edge Detection Chris Cacciatore, Tian Jiang, and Kerenne Paul University of Massachusetts Dartmouth Department of Mathematics Saeja Kim and Sigal Gottlieb. RBF’s and their Results. Original Image. - PowerPoint PPT Presentation

Transcript of Radial Basis Functions and Application in Edge Detection

Page 1: Radial Basis Functions and Application in Edge Detection

Radial Basis Functions and Application in Edge DetectionChris Cacciatore, Tian Jiang, and Kerenne Paul

University of Massachusetts Dartmouth Department of Mathematics Saeja Kim and Sigal Gottlieb

Abstract

This project focuses on the use of Radial Basis Functions in Edge Detection in two-dimensional images. We will be using a 2-D iterative RBF edge detection method. We will be varying the point distribution and shape parameter while also quantifying the effects of the accuracy of the edge detection on 2-D images. Furthermore, we study a variety of Radial Basis Functions and their accuracy in Edge Detection.

Radial Basis Functions (RBF’s)

Radial Basis Functions use the distance between two points on a given interval and epsilon (shape parameter) as variables. Three common types of RBF’s are Multi-quadric, Inverse Multi-quadric, and Gaussian.

Multi-quadric• =

Gaussian• Exp()

𝝐−𝒂𝒅𝒂𝒑𝒕𝒊𝒗𝒆𝒎𝒆𝒕𝒉𝒐𝒅This method changes the values of the shape parameters depending on the smoothness of f(x). Using this method allows the accuracy of the approximations to be solely determined on . The Main idea is that disappears only near the center of the discontinuity resulting in the basis functions near the discontinuity to become linear. This causes Gibbs oscillations not to appear in the approximation.

We have used three different radial basis functions and their results at different values of epsilon. Each function requires different values of epsilon to render a recognizable edge map.

Future Work

• Explore further into matrix involvement in edge detection

• Look into effects different parts of the code, TwoD_Example1, have on edge maps

• Test for necessary changes in epsilon for different sized images

RBF’s and their Results

Original image

M = zeros(N); MD = M; for ix = 1:N for iy = 1:N

M(ix,iy) = sqrt( (x(ix)-x(iy))^2 + (eps(iy))^2); if M(ix,iy) == 0 MD(ix,iy) = 0; else MD(ix,iy) = (x(ix) - x(iy))/M(ix,iy); end

Multi-quadric

Inverse Multi-quadric M = zeros(N); MD = M; for ix = 1:N for iy = 1:NM(ix,iy) = 1/sqrt( (x(ix)-x(iy))^2 + (eps(iy))^2); if M(ix,iy) == 0 MD(ix,iy) = 0; else MD(ix,iy) = -(x(ix) - x(iy))/sqrt( ((x(ix)-x(iy))^2 + (eps(iy))^2)^3); endGaussian M = zeros(N); MD = M; for ix = 1:N for iy = 1:NM(ix,iy) = exp(-((eps(iy))^2)*((x(ix)-x(iy))^2)); if M(ix,iy) == 0 MD(ix,iy) = 0; else MD(ix,iy) = -2*((eps(iy))^2)*(x(ix)-x(iy))*exp(-((eps(iy))^2)*(x(ix)-x(iy))^2); end

Multi-quadric

Gaussian

Inverse Multi-quadric

Original Image