Re: interpolation for a color image?

On May 29, 6:42 pm, aruzinsky <aruzin...@xxxxxxxxxxxxxxxxxxxx> wrote:
On May 29, 12:07 am, Martin Leese <ple...@xxxxxxxxxxxxxxxxxxxxxxxxxx>
wrote:





Harris wrote:
Original image sample:

1 2
3 4

Your 2-pass 1-D linear interpolation expands this box twice, once for X-direction and once for Y-
direction. The new pixels at position (area) 4 are interpolated using 3 and 2 respectively, that is using
distance 1. But the distance from 1, which should also be taken into account statistically, is at
distance sqrt(2). Even if you are using weighting factors, you end up with something else (weighted
average) other than true bilinear interpolation.

Sorry to interrupt, but bi-linear interpolation
is separable.  Always has been, always will be.
If you still doubt this then ten minutes with
pen and paper will convince you otherwise.

We now return you to your scheduled flame-fest.

--
Regards,
Martin Leese
E-mail: ple...@xxxxxxxxxxxxxxxxxxxxxxxxxx
Web:http://members.tripod.com/martin_leese/

Remember the saying, "All that is necessary for evil to succeed is
that good men do nothing."?   Replace "evil" with "stupidity".  You
are a good man.  Thank you.- Hide quoted text -

- Show quoted text -

You pretentious, pompous, condescending prat.

illywhacker;
.

Relevant Pages

  • Re: 3D distance problem
    ... > good idea of the average score on each variable and for each car brand. ... distance, so I'm not sure if we are on the same page in terms of a solution. ... There are many possible ways to do the weighting. ... prediction. ...
    (sci.stat.consult)
  • Re: interpolation for a color image?
    ... But the distance from 1, which should also be taken into account statistically, is at ... Even if you are using weighting factors, you end up with something else (weighted ... average) other than true bilinear interpolation. ... Remember the saying, "All that is necessary for evil to succeed is ...
    (sci.image.processing)
  • Re: interpolation for a color image?
    ... Harris wrote: ... But the distance from 1, which should also be taken into account statistically, is at distance sqrt. ... Even if you are using weighting factors, you end up with something else other than true bilinear interpolation. ... Martin Leese ...
    (sci.image.processing)
  • Re: Under what condition is Mahalanobis distance OPTIMAL?
    ... There are so many kinds of distance metrics. ... Is there a way one can show under what conditions the Mahalanobis ... better than Euclidean distance when the axes have different weighting, ...
    (sci.math)
  • Re: Under what condition is Mahalanobis distance OPTIMAL?
    ... There are so many kinds of distance metrics. ... Is there a way one can show under what conditions the Mahalanobis ... better than Euclidean distance when the axes have different weighting, ...
    (sci.stat.math)