0

I try to work the R transformation that, according with what is written in the scientific article "Recognition of Human Home Activities via Depth Silhouettes and R Transformation for Smart Homes", is computed in this way:

enter image description here

where

enter image description here

In Matlab I get the radon transform with:

[R,xp] = radon(Silhouette,theta);

Then, according with what is written this article, I compute the R transformation by adding,for every angle (that are the columns of R), the square of radon transform values (that, setting the column, are on the rows of R). But for each column this sum is the same. Where am I wrong?

Thanks for your support.

marc_s
  • 732,580
  • 175
  • 1,330
  • 1,459
Mario Lepore
  • 307
  • 2
  • 8
  • 18

1 Answers1

1

With the provided evidence I can only speculate on things that might be wrong.

If you get the same values after the integration (summation) you might want to check for bad values (or provide details/examples) on i) the way you define angle theta (range, degrees or radians etc.) ii) your Silhouette data.

Using MATLAB's example for theradon function, your profile should be extracted somewhat similar to: 

% input image
I = zeros(100,100); I(25:75, 25:75) = 1;

theta = 0:180;
[R,xp] = radon(I, theta);

Rt = sum(R.^2, 1);
figure; plot(theta, Rt);
gevang
  • 4,994
  • 25
  • 33
  • thanks man. The error was previously, when I transfer the Silhouette from my system written in c++ to matlab to do the R-transform. It was an input problem. – Mario Lepore Oct 18 '12 at 13:19