Algorithm to compute Basis images of an image

Click For Summary

Discussion Overview

The discussion revolves around the computation of basis images from a given image, exploring algorithms and code implementations related to image representation through basis functions, particularly in the context of Fourier analysis and other decomposition methods.

Discussion Character

  • Exploratory
  • Technical explanation
  • Mathematical reasoning

Main Points Raised

  • One participant notes that any image can be represented as a summation of basis images, similar to signals in Fourier analysis.
  • Another suggests searching for MPEG and JPEG code libraries to find relevant source code for computing basis images.
  • A different participant mentions that basis images can be derived from various function decompositions, highlighting the use of eigenvectors in computer vision and referencing SVD (Singular Value Decomposition) for image decomposition.
  • A participant shares their MATLAB code for computing the magnitude and phase spectrum of an image using the "fft2" function but expresses uncertainty about the next steps in their analysis.
  • One participant asks for clarification on the specific goals of the original poster regarding the computation of basis images.
  • Another participant interprets the original poster's intent as seeking frequency information from a two-dimensional signal and suggests methods for obtaining gray-scale information and reconstructing the signal.

Areas of Agreement / Disagreement

Participants express various approaches and suggestions for computing basis images, but there is no consensus on a specific algorithm or method. The discussion includes multiple competing views and interpretations of the original question.

Contextual Notes

Some limitations include the dependence on specific definitions of basis images and the unresolved nature of the next steps in the MATLAB implementation shared by one participant.

ramdas
Messages
78
Reaction score
0
I know from the Fourier Analysis
that any signal can be represented
as summation of elementary
signals i.e. basis functions
.Likewise,any image can be
represented as summation of Basis images.

Is there any available code, or
even an algorithm, that would
allow me to compute Basis images
of an image?
 
Technology news on Phys.org
Hey ramdas.

I think you should search for MPEG and JPEG code libraries and look at the source code for an open source project.

That should give you a good lot of code to start off with.
 
You have basis image from Fourier analysis and basis image from other function decomposition. Good basis images used in Computer vision are the eigen vector of the image. You can find explanation by looking to SVD, SVD compression, SVD image decomposition. This method give pertinent ortogonal basis to explain images.
 
@kroni ,@chiro Sir ,I have done basic programming in MATLAB. I have computed magnitude and phase spectrum using "fft2 "function in MATLAB.I am adding my code .But i don't know how to move forward
Code:
clc;
clear all;
close all;

i=imread('C:\Users\RK\Desktop\cameraman.gif');
%i=rgb2gray(i);
i=uint8(i);

figure,
subplot(1,3,1);imshow(i);
%title('\alpha^2 and X_1')
%title('\ite^-1000*|t|');
%title('e^{-1000|t|}');
title('\ite^{-1000 |t|}','Color','b')
%title('Cameraman Gray scale Image');

f1=fft2(i);
f2=log(1+fftshift(f1));

m=abs(f2);
subplot(1,3,2);
imshow(m,[]);
title('Magnitude Spectrum');

phase=angle(f2);
subplot(1,3,3);
imshow(phase,[]);
title('Phase Spectrum');
 
what do you want to obtain ?
 
I think he is trying to get the frequency information of a two-dimensional signal (since I noticed ff2 function there).

If you want to get the gray-scale information you should either do an average function of the RGB values or use a transform like they do in NTSC images.

Also try and plot the image based on re-constructing the signal with a specific bandwidth by creating a bitmap file and passing it to MATLAB for rendering.
 

Similar threads

Replies
4
Views
2K
  • · Replies 8 ·
Replies
8
Views
3K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 6 ·
Replies
6
Views
2K
Replies
13
Views
2K
  • · Replies 139 ·
5
Replies
139
Views
11K
Replies
14
Views
3K
  • · Replies 30 ·
2
Replies
30
Views
7K
  • · Replies 3 ·
Replies
3
Views
1K
  • · Replies 2 ·
Replies
2
Views
3K