Re: Calculating Variance in real-time

From: Jörg Neulist (ka-cj6_at_gmx.de)
Date: 06/27/04 

Date: Sun, 27 Jun 2004 11:13:42 +0200

Michael Newberry wrote:
> With all due respect, what you are saying is not going to work. There is not
> assumed any correlation between one image and the next. Any pixel of an
> image has no relation with the corresponding pixel in any other image.

How do you know? Depends strongly on the image source, I'd guess. A
typical application would be the data from a fixed camera, to gain a
denoised image.

Jörg

Relevant Pages

  • Re: Rate this analysis from 0 to 10
    ... I think the difference in correlation is significant. ... Real imagery has very high pixel correlation? ... Now a pixel isn't really square at the sensor level, ... But in the case you describe, sorry, random rotations ...
    (sci.image.processing)
  • Re: Bicubic interpolation
    ... alignment by running a root mean squares cross correlation. ... by sub-pixel correlation within half a pixel radius. ... but that would be very time consuming so, on balance, doing a full ...
    (sci.image.processing)
  • Re: Calculator
    ... I'm saying it is not, ... then fetch the pixel at those global coordinates; ... to use the old-style File ...
    (comp.lang.c)
  • Re: Rush for HD will create huge vacuum for non-HD viewing quality...
    ... drugs began to take hold. ... I remember "Steve Thackery" ... saying something like: ... pixel - 0.027mm on my display. ...
    (uk.tech.digital-tv)
  • Re: fft2 and ifft2 outputs
    ... contribution to the correlation even when there is no match. ... way to do this is simply set the first location of A3 and B2 to ... zero mean, that is to say it spends as much time with negative values ... from each pixel. ...
    (comp.soft-sys.matlab)