Getting Error In FFT filtering ?

sir i tried to make a gaussian filter.i use the steps given in the book(digital image processing using MATLAB,OF Gonzalez & Woods & Eddins )in the chapter3.i have done upto the step
[f,revertclass]=tofloat(f)
here 'f' is an image(of int8),here also given an error that"Undefined function or method 'tofloat' for input arguments of type 'uint8'."
PQ=paddedsize(size(f));
F=fft2(f,PQ(1),PQ(2));
then i use gaussian filter and get the transfer function.and then
H=ifftshift(H) for making center
then G=H.*F;
here i get an error which is "Error using ==> times Matrix dimensions must agree.
Error in ==> gaussian at 16
G=H.*F;"
please tell me how to overcome from this error

 Risposta accettata

Jordan Monthei
Jordan Monthei il 8 Mag 2013

0 voti

you are performing matrix multiplication when you use the '.*' command. The error you are receiving is telling you that the two "matrices" that you are trying to multiply are not of the same dimensions.
I assume you are trying to multiply numbers and not matrices and would recommend using '*' and not '.*'

3 Commenti

[f,revertclass]=tofloat(f)
This line is reading the value of 'f' which is returned from the function 'tofloat'. Is this function being declared in your program prior to being called?
Jordan, what you said about multiplication is exactly the opposite of what it is. * is matrix multiplication and .* is element-by-element multiplication.
I stand corrected. good catch.

Accedi per commentare.

Più risposte (1)

Make sure H and F are the same size.
size(H)
size(f)
size(PQ)
size(F)
What does all that report?

Categorie

Prodotti

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by