Login  Register

Re: RGB to CMYK

Posted by Kateo on Jun 01, 2010; 6:07pm
URL: http://imagej.273.s1.nabble.com/RGB-to-CMYK-tp3688100p3688104.html

HI Stephan,

Thank you very much for your reply! I appreciate your insight, and  
agree that perhaps I don't need to convert my image to CMYK.

I have tried using the color deconvoluter. However, I do not believe  
it will work with the type of images I am looking at. We stained them  
with picrosirius red to enhance the bifringency of collagen. We then  
looked at the slides under polarized microscopy and from that took  
pictures of the light that passed through the polarizing plates and  
collagen. So since the color in my pictures is not due to a stain but  
due to light passing through collagen of different thickness, I think  
this plug in is of no help, unfortunately.

I am still trying to figure out how to get a % of red/orange and  
green/yellow that are in my pictures since mature collagen in  
red/orange and immature collagen is yellow/green.

If you have any further ideas or advice, please pass it on! I'm green  
behind the ears and am slowly getting discouraged.

Thanks very much,

Kate


Quoting "Stephan Saalfeld [via ImageJ]"  
<[hidden email]>:

>
>
>
> In the trivial case, that is when the intensity value stored for each
> color channel is linear, grey is ( red + green + blue ) / 3 and no color
> profiles are in the way, the conversion is quite easy.  Let r, g and b
> be red, green and blue in unsigned byte range [0...255] and calculate c,
> m, y and k as floating point in range [0...1]:
>
> c = 1.0 - r/255.0
> m = 1.0 - g/255.0
> y = 1.0 - b/255.0
>
> That's CMY from which CMYK is derived by extracting the grey component
> into a fourth channel and scaling the remaining color channels
> respectively:
>
> k = min( c, y, k )
> if ( k == 1.0 )
>   c = m = y = 0
> else
>   s = 1.0 - k
>   c = ( c - k ) / s
>   m = ( m - k ) / s
>   y = ( y - k ) / s
>
> For the application, you have in mind, I do not think that conversion
> into CMYK is of any help.  It does not add any information to your data.
> Better look at Gabriel Landini's Colour Deconvolution plugin and
> directly estimate the percentage of the colors you're looking for:
>
> http://www.dentistry.bham.ac.uk/landinig/software/cdeconv/cdeconv.html
>
> Best,
> Stephan
>
>
>
> On Thu, 2010-05-13 at 06:44 -0700, Kateo wrote:
>> HI Duane,
>> I was wondering if you could expand on how to convert RGB--> CMYK. I have no
>> clue about these things, but I found a paper that before analyzing the
>> colours in histological sections performed colour separation into CMYK
>> before dertermining the % of colours in their section. I want to determine
>> the % of orange/red vs. green colour in my sections stained with picrosirius
>> red and viewed under a polarized microscope. Anyway, I would appreciate your
>> guidance... or anyone else's!
>> Thanks,
>>
>> Kate
>
>
> ______________________________________
> View message @  
> http://imagej.588099.n2.nabble.com/RGB-to-CMYK-tp631553p5047771.html
>
> To unsubscribe from Re: RGB to CMYK, click  
> http://imagej.588099.n2.nabble.com/subscriptions/Unsubscribe.jtp?code=a2F0ZS5zYXVrc0B1dG9yb250by5jYXw1MDQ2MTczfC02OTIzNTgzNjg=
>