Re: usage of imread.m in octave

[Daniel J Sebald]

From what I've seen, only gray scale and color palettes (i.e., 
"colormap") currently work in Octave. To show a color image, you need 
independent primary color components, e.g., RGB. So first, you can't use

X = imread('testimage.jpg');

because X can't have RGB. (Unless it were a 3 dimensional matrix, which 
I think it isn't.) So you need something like the RGB variant:

[r,g,b]= imread( fname );

Then you can do an

imshow(r,g,b);

But, currently here is what imshow() does for the RGB

elseif (nargin == 3)
[a1, a2] = rgb2ind (a1, a2, a3);
colormap (a2);
endif

i.e., converts to an indexed image. I believe in general that can only 
be an approximation.

Dan



Ole Jacob Hagen wrote:

> Hi.
>
> Octave-Forge has imread.m as part of the signal processing module inside.
> I am using imread like this:
> X = imread('testimage.jpg');
> The testimage.jpg can be found here
> http://cvs.sourceforge.net/viewcvs.py/oplot/oplot/examples/
>
> And imshow(X), will then give me a black-white image, but it's supposed
> to be a colour image. I'd like to have colours.
> Does imshow or image work with colour?
>
> I would like to have working example script, if anyone has a working 
> example...if it's possible....
>
>
> Why doesn't this works? Where does it fails, what is wrong, and does
> someone have a fix for this?
>
> Oplot-developers have implemented the same thing, but the image popups
> in Oplot instead of ImageMagic. :-)
>
> We can do:
>
> Line, Image, Axes, Figure.
> Soon Patch and Surface.
>
> Then we are heading for UIControl...
>
>
>
> Cheers,
>
> Ole J.

--

Dan Sebald
email: daniel DOT sebald AT ieee DOT org
URL: http://acer-access DOT com/~dsebald AT acer-access DOT com/