Digital Noise Links Image to Camera

This message is not archived
 >> Stay informed about: Digital Noise Links Image to Camera 

fishfry writes:

> Interesting! They can use the noise in an image to id the exact camera
> the image came from.
>
> http://dsc.discovery.com/news/briefs/20060501/digitalnoise_tec.html

No surprise here. It's roughly the same as identifying a gun by the
marks it makes on a bullet, and of comparable utility and reliability.

--
Transpose mxsmanic and gmail to reach me by e-mail.
 >> Stay informed about: Digital Noise Links Image to Camera 

fishfry wrote:
> Interesting! They can use the noise in an image to id the
> exact camera the image came from.
>
> http://dsc.discovery.com/news/briefs/20060501/digitalnoise_tec.html

I've heard of similar research, and I always think about
working the other side to beat these techniques. If they
can profile my camera, so can I. If there's a consistent
patter, I ought to be able to correct for it. Then if I
have access to a bunch of photos from some other camera,
I ought to be able to develop a profile and add that
camera's noise pattern in.

How odd: the review sample lent to bobs-digital-cameras.com
also photographed all this porn.


--
--Bryan
 >> Stay informed about: Digital Noise Links Image to Camera 

fishfry wrote:
> Interesting! They can use the noise in an image to id the exact camera
> the image came from.
>
>
> http://dsc.discovery.com/news/briefs/20060501/digitalnoise_tec.html

My remarks apply only to the news item. I haven't read the
eoriginal publication.

By technical definition, noise is random. Therefore, noise
in a camera does not repeat from photo to photo. There are
irregularities in the sensor response that, in a good
camera, are mapped and corrected by software. Some
irregularities may not be corrected so well that they don't
affect the image. I'd be surprised if, except in exceptional
cases, the residual irregularites can be separated from the
image and the true noise.
 >> Stay informed about: Digital Noise Links Image to Camera 

Marvin <physchem.RemoveThis@cloud9.net> wrote:
>fishfry wrote:
>> Interesting! They can use the noise in an image to id the
>> exact camera the image came
>> from. http://dsc.discovery.com/news/briefs/20060501/digitalnoise_tec.html
>
>My remarks apply only to the news item. I haven't read the
>eoriginal publication.
>
>By technical definition, noise is random.

A false premise that invalidates your entire statement.

>Therefore, noise in a
>camera does not repeat from photo to photo. There are
>irregularities in the sensor response that, in a good camera,
>are mapped and corrected by software. Some irregularities may
>not be corrected so well that they don't affect the image. I'd
>be surprised if, except in exceptional cases, the residual
>irregularites can be separated from the image and the true noise.

--
Floyd L. Davidson <http://www.apaflo.com/floyd_davidson>
Ukpeagvik (Barrow, Alaska) floyd.RemoveThis@apaflo.com
 >> Stay informed about: Digital Noise Links Image to Camera 

Marvin writes:

> By technical definition, noise is random. Therefore, noise
> in a camera does not repeat from photo to photo.

Noise can be consistent from one photo to the next when it is due to
random, permanent defects in the sensor or in other parts of the
camera (usually the sensor, though).

> There are irregularities in the sensor response that, in a good
> camera, are mapped and corrected by software.

They cannot be corrected, they can only be masked. Masking them won't
prevent them from being used for identification.

> Some irregularities may not be corrected so well that they don't
> affect the image.

That isn't possible. Defects always degrade the image.

> I'd be surprised if, except in exceptional
> cases, the residual irregularites can be separated from the
> image and the true noise.

I'd be surprised if they couldn't. It's not rocket science.

--
Transpose mxsmanic and gmail to reach me by e-mail.
 >> Stay informed about: Digital Noise Links Image to Camera