Fwd: Colour Quantification

[S. Pagakis (nimr)]

Hello

We would like to quantify the "amount" of a colour.

We have two RGB pictures in which an area appears brown and another  
one lighter brown and our experiment will produce a series of "brown"  
intensities in between.

We know the ImageJ tools which measure RGB intensity, but expressing  
the two areas as RGB intensities does not produce something  
"meaningful" which can be graphed.

Has anyone used a specific methodology (or has a plugin) which can  
quantify this in another way?  Possibly converting the RGB into  
another colour space?

Regards

Samatis Pagakis


*********************************
Stamatis Pagakis Ph.D.
Biological Imaging Unit
Biomedical Research Foundation, Academy of Athens
Soranou Efessiou 4, Athens 115 27 - Greece
M:     [phone]
W:     [phone]
FAX: [phone]
http://www.bioacademy.gr/Faculty/core.php?cr=17

[karo03]

Hi,
I have used Lch color space which is calculated from Lab color space  
(c=sqrt(a^2 + b^2), h=arctan(a,b)). In this color space L is the  
luminance, h the hue (an angle) and c the cromaticy. Color brown might  
be always a problem, however, possibly the hue and the cromaticy help  
to order the colors.
Regards
Karsten

Am 07.04.2009 um 12:32 schrieb S. Pagakis:

> Hello
>
> We would like to quantify the "amount" of a colour.
>
> We have two RGB pictures in which an area appears brown and another  
> one lighter brown and our experiment will produce a series of  
> "brown" intensities in between.
>
> We know the ImageJ tools which measure RGB intensity, but expressing  
> the two areas as RGB intensities does not produce something  
> "meaningful" which can be graphed.
>
> Has anyone used a specific methodology (or has a plugin) which can  
> quantify this in another way?  Possibly converting the RGB into  
> another colour space?
>
> Regards
>
> Samatis Pagakis
>
>
> *********************************
> Stamatis Pagakis Ph.D.
> Biological Imaging Unit
> Biomedical Research Foundation, Academy of Athens
> Soranou Efessiou 4, Athens 115 27 - Greece
> M:     [phone]
> W:     [phone]
> FAX: [phone]
> http://www.bioacademy.gr/Faculty/core.php?cr=17

Karsten
[hidden email]

[Stephan Saalfeld]

Hi Samatis,

if there is only "brown" in the image.  The intensity itself is the
"amount of color".  If you have more colors and just the "brown" has
meaning for you, you would have to make "brown" an explicit axis of your
color space which is just a linear transformation.  Since you are not
interested in the other components of this color space, you can estimate
the "amount of brown" of an rgb-color c by

amount_of_brown =
  c.red / brown.red +
  c.green / brown.green +
  c.blue / brown.blue

Blame me an idiot if this is wrong ;)

Best regards,
Stephan


On Tue, 2009-04-07 at 13:32 +0300, S. Pagakis wrote:

> Hello
>
> We would like to quantify the "amount" of a colour.
>
> We have two RGB pictures in which an area appears brown and another  
> one lighter brown and our experiment will produce a series of
> "brown"  
> intensities in between.
>
> We know the ImageJ tools which measure RGB intensity, but expressing  
> the two areas as RGB intensities does not produce something  
> "meaningful" which can be graphed.
>
> Has anyone used a specific methodology (or has a plugin) which can  
> quantify this in another way?  Possibly converting the RGB into  
> another colour space?
>
> Regards
>
> Samatis Pagakis
>
>
> *********************************
> Stamatis Pagakis Ph.D.
> Biological Imaging Unit
> Biomedical Research Foundation, Academy of Athens
> Soranou Efessiou 4, Athens 115 27 - Greece
> M:     [phone]
> W:     [phone]
> FAX: [phone]
> http://www.bioacademy.gr/Faculty/core.php?cr=17

[Stephan Saalfeld]

Ouch!  I have to blame me an idiot ;)  The first part sounds reasonable,
but the solution is wrong.

Will take me some more minutes to come up with something ...

Best regards,
Stephan


On Tue, 2009-04-07 at 13:27 +0200, Stephan Saalfeld wrote:

> Hi Samatis,
>
> if there is only "brown" in the image.  The intensity itself is the
> "amount of color".  If you have more colors and just the "brown" has
> meaning for you, you would have to make "brown" an explicit axis of
> your
> color space which is just a linear transformation.  Since you are not
> interested in the other components of this color space, you can
> estimate
> the "amount of brown" of an rgb-color c by
>
> amount_of_brown =
>   c.red / brown.red +
>   c.green / brown.green +
>   c.blue / brown.blue
>
> Blame me an idiot if this is wrong ;)
>
> Best regards,
> Stephan
>
>
> On Tue, 2009-04-07 at 13:32 +0300, S. Pagakis wrote:
> > Hello
> >
> > We would like to quantify the "amount" of a colour.
> >
> > We have two RGB pictures in which an area appears brown and
> another  
> > one lighter brown and our experiment will produce a series of
> > "brown"  
> > intensities in between.
> >
> > We know the ImageJ tools which measure RGB intensity, but
> expressing  
> > the two areas as RGB intensities does not produce something  
> > "meaningful" which can be graphed.
> >
> > Has anyone used a specific methodology (or has a plugin) which can  
> > quantify this in another way?  Possibly converting the RGB into  
> > another colour space?
> >
> > Regards
> >
> > Samatis Pagakis
> >
> >
> > *********************************
> > Stamatis Pagakis Ph.D.
> > Biological Imaging Unit
> > Biomedical Research Foundation, Academy of Athens
> > Soranou Efessiou 4, Athens 115 27 - Greece
> > M:     [phone]
> > W:     [phone]
> > FAX: [phone]
> > http://www.bioacademy.gr/Faculty/core.php?cr=17