On 11 April 2012 16:24, Carnë Draug <carandraug+...@gmail.com> wrote: > On 7 April 2012 12:55, Martin Helm <mar...@mhelm.de> wrote: >> Am 06.04.2012 22:47, schrieb Alexander Görtz: >> >>> Hi, >>> >>> I'm new at this list, so first hello to everyone and keep up the good >>> work. >>> >>> But now to my 'problem'. I'm using the im2double function for some >>> grayscale >>> images which worked fine until I came across a 1bit black and white image. >>> I >>> got the error that the input of im2double must be an image, which doesn't >>> make >>> that much sens as the input actually is an image. So I would suggest to >>> change >>> the error message that the input hast to be a 8, 16 uint or a double >>> image. >>> Or even better make im2double work with class(img) = logical what should >>> be >>> fairly easy as the only thing you need to return is double(img) as the >>> image >>> already is in the range [0, 1]. >>> >>> Cheers, >>> >>> Alexander >>> >> Can you try the attached one? If it works as expected I will send a patch >> for the svn sources. > > I have made some further changes to im2double to add support for > indexed images. Also increased matlab compatibility and now accepts > input of double precision (returns the same) instead of an error.
I have expanded the changes to the others im2"class" functions and added 2 simple implementations of new ones (im2single and im2int16). Carnë ------------------------------------------------------------------------------ Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev _______________________________________________ Octave-dev mailing list Octave-dev@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/octave-dev
