Thread Subject: how to transform the power spectra image into spectrum diagram?

Subject: how to transform the power spectra image into spectrum diagram?

From: islinaismail@gmail.com

Date: 9 Apr, 2009 10:07:17

Message: 1 of 7

hye...

how to transform the power spectra image (through fft2 process) into
spectrum diagram?is anyone know the codes in matlab?
thanks...

Subject: how to transform the power spectra image into spectrum diagram?

From: Nasser Abbasi

Date: 9 Apr, 2009 11:08:21

Message: 2 of 7


<islinaismail@gmail.com> wrote in message
news:e86ceb09-d620-496d-a2a3-fa987954c5d9@k8g2000yqn.googlegroups.com...
> hye...
>
> how to transform the power spectra image (through fft2 process) into
> spectrum diagram?is anyone know the codes in matlab?
> thanks...

The power spectrum is the square of the magnitude of the spectrum?

So, just take the square root of the PSD you'll get |X(w)|

--Nasser

Subject: how to transform the power spectra image into spectrum diagram?

From: islinaismail@gmail.com

Date: 9 Apr, 2009 11:52:32

Message: 3 of 7

I want to ask...is it possible to differentiate 2 different images by
using the fft2 function in matlab?
If i'm not mistaken, fft2 converts the image into cos and sin
function..Is it possible to differentiate the images (power spectrum
image) by looking only at its phase and magnitude diagram?thanks..

Subject: how to transform the power spectra image into spectrum diagram?

From: Nasser Abbasi

Date: 9 Apr, 2009 12:08:33

Message: 4 of 7


<islinaismail@gmail.com> wrote in message
news:12b90af4-28c5-42f5-bf54-4234749f1364@r28g2000vbp.googlegroups.com...
>I want to ask...is it possible to differentiate 2 different images by
> using the fft2 function in matlab?
> If i'm not mistaken, fft2 converts the image into cos and sin
> function..Is it possible to differentiate the images (power spectrum
> image) by looking only at its phase and magnitude diagram?thanks..

I think you are mixing terms a little here.

PSD is not the spectrum. PSD has no phase information in it, only magnitude
information of the spectrum. PSD=|X(w)|^2.

Therefore, different images can give the same PSD, it is not unique.

You need both phase and magnitude. If you are looking at spectrum only,
then looking at the magnitude and phase of each image spectrum, you can tell
if these are from same or different image if both magnitude *and* phase data
is exactly the same or not.

--Nasser

Subject: how to transform the power spectra image into spectrum diagram?

From: islinaismail@gmail.com

Date: 9 Apr, 2009 12:49:46

Message: 5 of 7

Oh...i see....but,let say if i have 2 images that absolutely same but
was captured in the different angle/direction.Does fft2 can detect it
as the same image source not? thanks...

Subject: how to transform the power spectra image into spectrum diagram?

From: Dave Robinson

Date: 9 Apr, 2009 13:19:01

Message: 6 of 7

islinaismail@gmail.com wrote in message <a0e8c843-4124-4c2c-8bbf-ef1fb8aeb646@t11g2000vbc.googlegroups.com>...
> Oh...i see....but,let say if i have 2 images that absolutely same but
> was captured in the different angle/direction.Does fft2 can detect it
> as the same image source not? thanks...

No - they will be detected as different images. The 2D transform gets rotated by the same amount as the original image subject, and thus will not match.

Regards

Dave Robinson

Subject: how to transform the power spectra image into spectrum diagram?

From: islinaismail@gmail.com

Date: 9 Apr, 2009 18:11:21

Message: 7 of 7

Is that any method that i can do to detect two images that is
absolutely same as the same image source eventhough it was captured in
different angle or direction?plz...i need your assist since i have to
finish my final year project regarding to the authenticity of fabric
image .tq...

Tags for this Thread

Add a New Tag:

Separated by commas
Ex.: root locus, bode

What are tags?

A tag is like a keyword or category label associated with each thread. Tags make it easier for you to find threads of interest.

Anyone can tag a thread. Tags are public and visible to everyone.

rssFeed for this Thread

Contact us at files@mathworks.com