Help with mahotas.label ( 6 months on trying...)

[Eduardo Munizaga]

Hello friends, I spent six months trying and still can not find solution. I
need help please

I'm trying to count objects with mahotas.label algorithm. All is well when
the sample image is clean. but when there is any unwanted object, I can not
make it work.

Is it possible to apply the algorithm only in a range of RGB color desired?

Deputy experiment.

<Loading Image...>


Link:
http://nbviewer.ipython.org/url/www.pokytools.cl/notebook/problema.ipynb


Thank you very much for your help

Eduardo Munizaga
From Chile
(Sorry for my bad English)

[Luis Pedro Coelho]

Post by Eduardo Munizaga
I'm trying to count objects with mahotas.label algorithm. All is well
when the sample image is clean. but when there is any unwanted object, I
can not make it work.
Is it possible to apply the algorithm only in a range of RGB color desired?

I would approach this differently. I would filter the unwanted objects
afterwards.

Now you need some criterion for object filtering. You can try to isolate
each object and compare its colour profile. For this, convert to HSV or
LAB scale and try a few rules like:

If 50% or more of the pixels in an object are within distance D of
colour C, reject this object.

HTH
Luis
--
Luis Pedro Coelho | EMBL | http://luispedro.org
http://metarabbit.wordpress.com

Post by Eduardo Munizaga
Deputy experiment.
<https://lh4.googleusercontent.com/-YjzbwmTal4M/U0XxCriM_OI/AAAAAAAAAUE/te7VA9Ouecs/s1600/experimento.png>
http://nbviewer.ipython.org/url/www.pokytools.cl/notebook/problema.ipynb
<http://nbviewer.ipython.org/url/www.pokytools.cl/notebook/problema.ipynb>
Thank you very much for your help
Eduardo Munizaga
From Chile
(Sorry for my bad English)
--
You received this message because you are subscribed to the Google
Groups "pythonvision" group.
To unsubscribe from this group and stop receiving emails from it, send
For more options, visit https://groups.google.com/d/optout.

[Eduardo Munizaga]

Ok thanks, I'll have to do the experiments and commented how it works
Regards
Eduardo

Post by Eduardo Munizaga

Post by Eduardo Munizaga
I'm trying to count objects with mahotas.label algorithm. All is well
when the sample image is clean. but when there is any unwanted object, I
can not make it work.
Is it possible to apply the algorithm only in a range of RGB color

desired?
I would approach this differently. I would filter the unwanted objects
afterwards.
Now you need some criterion for object filtering. You can try to isolate
each object and compare its colour profile. For this, convert to HSV or
If 50% or more of the pixels in an object are within distance D of
colour C, reject this object.
HTH
Luis
--
Luis Pedro Coelho | EMBL | http://luispedro.org
http://metarabbit.wordpress.com

Post by Eduardo Munizaga
Deputy experiment.
<

https://lh4.googleusercontent.com/-YjzbwmTal4M/U0XxCriM_OI/AAAAAAAAAUE/te7VA9Ouecs/s1600/experimento.png>

Post by Eduardo Munizaga
http://nbviewer.ipython.org/url/www.pokytools.cl/notebook/problema.ipynb
<

http://nbviewer.ipython.org/url/www.pokytools.cl/notebook/problema.ipynb>

Post by Eduardo Munizaga
Thank you very much for your help
Eduardo Munizaga
From Chile
(Sorry for my bad English)