| Calculating colocalization of three colours [message #66692] |
Thu, 04 June 2009 09:42  |
cgguido
Messages: 195
Registered: August 2005
|
Senior Member |
|
|
Hi all,
say I have two images, r(ed) and g(reen), and I want to know how
colocalized these colours are. I do c1=correlate(r,g). if c is close
to 1 then there is a lot of colocalization, if c~0 then there is none,
if c~-1 then some joker just gave me the same image twice, inverting
one of the copies! (usually, one calculates c on a ROI...)
I am trying to figure out how to do something similar when a b(lue)
image is added to the mix. I could do them pairwise, but that means
for each set, I would end up with three numbers...
Any ideas?
Many thanks,
Gianguido
|
|
|
|
| Re: Calculating colocalization of three colours [message #66749 is a reply to message #66692] |
Wed, 10 June 2009 12:50  |
cgguido
Messages: 195
Registered: August 2005
|
Senior Member |
|
|
On Jun 10, 2:42 pm, gianguido.cia...@gmail.com wrote:
> Grazie for the reply Paolo... ;-)
>
> I guess if you consider a triangle with side lengths C_ab, C_ac, and
> C_bc, where these are the correlation coefficients, you can describe
> the triangle with only two numbers if you put one apex at the origin
> and one side along an axis... hmmm.
>
> Anyway thanks again,
> Gianguido
No no, still need 3 numbers! doh!
G
|
|
|
|
| Re: Calculating colocalization of three colours [message #66750 is a reply to message #66692] |
Wed, 10 June 2009 12:42 |
cgguido
Messages: 195
Registered: August 2005
|
Senior Member |
|
|
Grazie for the reply Paolo... ;-)
I guess if you consider a triangle with side lengths C_ab, C_ac, and
C_bc, where these are the correlation coefficients, you can describe
the triangle with only two numbers if you put one apex at the origin
and one side along an axis... hmmm.
Anyway thanks again,
Gianguido
On Jun 10, 10:29 am, Paolo <pgri...@gmail.com> wrote:
> On Jun 4, 12:42 pm, Gianguido Cianci <gianguido.cia...@gmail.com>
> wrote:
>
>
>
>> Hi all,
>
>> say I have two images, r(ed) and g(reen), and I want to know how
>> colocalized these colours are. I do c1=correlate(r,g). if c is close
>> to 1 then there is a lot of colocalization, if c~0 then there is none,
>> if c~-1 then some joker just gave me the same image twice, inverting
>> one of the copies! (usually, one calculates c on a ROI...)
>
>> I am trying to figure out how to do something similar when a b(lue)
>> image is added to the mix. I could do them pairwise, but that means
>> for each set, I would end up with three numbers...
>
>> Any ideas?
>
>> Many thanks,
>> Gianguido
>
> Well, considering that you have 3 possibilities:
>
> - all 3 correlated
> - 2 correlated, one not
> - none are correlated
>
> I don't think anything less than 3 numbers would be enough anyway...
>
> Ciao,
> Paolo
>
> Ciao,
> Paolo
|
|
|
|
| Re: Calculating colocalization of three colours [message #66751 is a reply to message #66692] |
Wed, 10 June 2009 08:29 |
pgrigis
Messages: 436
Registered: September 2007
|
Senior Member |
|
|
On Jun 4, 12:42 pm, Gianguido Cianci <gianguido.cia...@gmail.com>
wrote:
> Hi all,
>
> say I have two images, r(ed) and g(reen), and I want to know how
> colocalized these colours are. I do c1=correlate(r,g). if c is close
> to 1 then there is a lot of colocalization, if c~0 then there is none,
> if c~-1 then some joker just gave me the same image twice, inverting
> one of the copies! (usually, one calculates c on a ROI...)
>
> I am trying to figure out how to do something similar when a b(lue)
> image is added to the mix. I could do them pairwise, but that means
> for each set, I would end up with three numbers...
>
> Any ideas?
>
> Many thanks,
> Gianguido
Well, considering that you have 3 possibilities:
- all 3 correlated
- 2 correlated, one not
- none are correlated
I don't think anything less than 3 numbers would be enough anyway...
Ciao,
Paolo
Ciao,
Paolo
|
|
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
