Scale a vector

[Ben Voigt [C++ MVP]]

I don't have enough experience to know the actual answer, but both people

who have told me I'm wrong do not appear to have been paying close enough
attention to my actual suggestion. It would be nice to see a reply that
answers the question while being specific enough to be clear that the
answer takes into account my actual suggestion.

In this particular case, it wasn't clear that your actual suggestion spoke
of creating a new image with the values in question, because you mentioned
copying pixels. I didn't see any statement that the "vector" being
resampled contained color values, it seemed more likely to contain some
position information.

Yes, you could probably use the built-in .NET image processing functions on
the values -- if they fit within the range of a single color (which the OP
did guarantee). But it really sounded as if you were suggesting resizing
the image before performing the feature identification.

 

[Peter Duniho]

[...]
Yes, you could probably use the built-in .NET image processing functions on
the values -- if they fit within the range of a single color (which the OP
did guarantee). But it really sounded as if you were suggesting resizing
the image before performing the feature identification.

Sorry for the confusion. Please see option #2 in this post:
http://groups.google.com/group/micr...ages.csharp/msg/d881737c991de5ea?dmode=source

It may have been unreasonable for me to expect people to have that post
in mind when I asked my question, but that was in fact the scenario to
which I was referring.

I took it as obvious that scaling the entire image would be less
efficient (and even said so in the above-referenced post) and just
assumed (apparently incorrectly) that it would be understood that by
suggesting the use of the built-in functionality that does bi-cubic
interpolation rather than writing a from-scratch implementation, I did
not intend to do it in the inefficient way, but rather only to
interpolate the data that actually requires interpolation.

Hope that clears things up. I'm still at this point left thinking that
using the built-in functionality is likely to be fine and would provide
for a simple, error-free implementation as compared to coding up one's
own cubic interpolation.

Pete