Posted by
Burger Wilhelm on
Mar 11, 2009; 9:44pm
URL: http://imagej.273.s1.nabble.com/Zooming-and-smoothing-tp3693324p3693339.html
Oh sure, but this would be in line with the current implementation and would not break existing code. The switch would not be global but "private" to the individual ImageProcessor (with set-method etc.). Easy and clean enough for my taste.
Wilhelm
________________________________
Von: ImageJ Interest Group im Auftrag von Johannes Schindelin
Gesendet: Mi 11.03.2009 22:08
An:
[hidden email]
Betreff: Re: Zooming new problem
Hi,
On Wed, 11 Mar 2009, Burger Wilhelm wrote:
> as you mention, we have listed a number of different interpolation
> methods in the source section of our website (imagingbook.com),
> including cubic and other spline methods. I particularly recommend the
> Catmull-Rom formulation for good results. I can also send you code for a
> Lanczos interpolator, which has no visible advantage though.
>
> All these methods are pretty standard and implemented by stand-alone
> code that could be copied directly into ImageProcessor's
> getInterpolatedPixel() method and selected using a suitable mode switch.
Usually I would not say anything, but since we discussed clean design
earlier: I think it would not be appropriate to put this method into
ImageProcessor (think MVC), at least not with a global switch as to which
interpolation method is actually used.
Ciao,
Dscho