[Insight-users] LabelObject OCR with numbers

[Gaëtan Lehmann]

Le 27 sept. 10 à 17:37, Christian Werner a écrit :

> Weird, weird, weird. This really worked! How can a Centroids X- 
> position differ between two numbers that are _vertically_ mirrored?

because they are not mirrored but rotated by 180° :-)
I guess you should also see something of that kind on the Y axis BTW.

>
> Here is what I get:
>
> 0:  Size: 18 - Elongation: 1.44449 - Flatness: 1.44449 - Centroid:  
> 5,7 - dif: 2
> 1:  Size: 10 - Elongation: 3.65503 - Flatness: 3.65503 - Centroid:  
> 20,7 - dif: 1
> 2:  Size: 15 - Elongation: 2.06489 - Flatness: 2.06489 - Centroid:  
> 35,7 - dif: 2
> 3:  Size: 15 - Elongation: 1.86812 - Flatness: 1.86812 - Centroid:  
> 50,7 - dif: 2
> 4:  Size: 16 - Elongation: 1.63288 - Flatness: 1.63288 - Centroid:  
> 65,7 - dif: 2
> 5:  Size: 17 - Elongation: 1.7793 - Flatness: 1.7793 - Centroid:  
> 81,7 - dif: 2
> 6:  Size: 19 - Elongation: 1.54404 - Flatness: 1.54404 - Centroid:  
> 95,7 - dif: 1
> 7:  Size: 12 - Elongation: 2.50164 - Flatness: 2.50164 - Centroid:  
> 108,8 - dif: 1
> 8:  Size: 19 - Elongation: 1.53748 - Flatness: 1.53748 - Centroid:  
> 123,7 - dif: 2
> 9:  Size: 19 - Elongation: 1.54404 - Flatness: 1.54404 - Centroid:  
> 138,7 - dif: 2

Note that Elongation and Flatness are the same in 2D - they differ  
only in 3D (and more).

Gaëtan

>
> ...but I don't get it. :) Anyway, thanks alot Sergio and Juan!
>
>
> Best regards,
> Christian
>
>
> Am 27.09.2010 17:25, schrieb Juan Vidal:
>>
>> What about using Sergio's idea in a "per-number" basis?
>>
>> You could substract the x-positoin of the leftmost pixel of the  
>> number to its CentroidX, and maybe that result will help you to  
>> differenciate both numbers.
>>
>> Good luck!
>>
>>
>>
>> 2010/9/27 Christian Werner <christian.werner at rwth-aachen.de>
>> Thanks for the response, but the CentroidX already changes with the  
>> position of the number. So if I am analyzing a sequence of numbers  
>> "146998" the CentroidX will be different than in the sequence  
>> "6138549".
>>
>> This is really funny, who is responsible for that? Why use 9  
>> symbols to represent 10 digits? :)
>>
>>
>>
>> Am 27.09.2010 17:02, schrieb Sergio Vera:
>>> Hello
>>> Maybe it's a crazy/stupid/already tried idea but, what about  
>>> adding CentroidX to the list of features? I would say that '6'  
>>> should have the centroid more to the left and '9' more to the  
>>> right...
>>>
>>> HTH
>>>
>>> Sergio
>>>
>>> On Mon, Sep 27, 2010 at 4:44 PM, Christian Werner <christian.werner at rwth-aachen.de 
>>> > wrote:
>>>  Hi!
>>>
>>> I am doing a nice little OCR for some numbers. See the attached  
>>> image. I can identify each number as an object. When I look at  
>>> each object's attributes, I see a fatal similarity between the  
>>> number '6' and '9':
>>>
>>> 0: Size: 18 - Elongation: 1.44449 - Flatness: 1.44449 - CentroidY: 6
>>> 1: Size: 10 - Elongation: 3.65503 - Flatness: 3.65503 - CentroidY: 6
>>> 2: Size: 15 - Elongation: 2.06489 - Flatness: 2.06489 - CentroidY: 6
>>> 3: Size: 15 - Elongation: 1.86812 - Flatness: 1.86812 - CentroidY: 6
>>> 4: Size: 16 - Elongation: 1.63288 - Flatness: 1.63288 - CentroidY: 6
>>> 5: Size: 17 - Elongation: 1.7793 - Flatness: 1.7793 - CentroidY: 6
>>> 6: Size: 19 - Elongation: 1.54404 - Flatness: 1.54404 - CentroidY: 6
>>> 7: Size: 12 - Elongation: 2.50164 - Flatness: 2.50164 - CentroidY: 7
>>> 8: Size: 19 - Elongation: 1.53748 - Flatness: 1.53748 - CentroidY: 6
>>> 9: Size: 19 - Elongation: 1.54404 - Flatness: 1.54404 - CentroidY: 6
>>>
>>> Which means, that I can't distinguish '6' from a '9'. Does anybody  
>>> have any idea how I could do that?
>>>
>>> Best regards,
>>> Christian
>>>
>>> _____________________________________
>>> Powered by www.kitware.com
>>>
>>> Visit other Kitware open-source projects at
>>> http://www.kitware.com/opensource/opensource.html
>>>
>>> Kitware offers ITK Training Courses, for more information visit:
>>> http://www.kitware.com/products/protraining.html
>>>
>>> Please keep messages on-topic and check the ITK FAQ at:
>>> http://www.itk.org/Wiki/ITK_FAQ
>>>
>>> Follow this link to subscribe/unsubscribe:
>>> http://www.itk.org/mailman/listinfo/insight-users
>>>
>>>
>>>
>>>
>>> -- 
>>> Sergio Vera
>>>
>>>  Alma IT Systems
>>>  C/ Vilana, 4B, 4º 1ª
>>>  08022 Barcelona
>>>  T. (+34) 932 380 592
>>>  www.alma3d.com
>>
>>
>> _____________________________________
>> Powered by www.kitware.com
>>
>> Visit other Kitware open-source projects at
>> http://www.kitware.com/opensource/opensource.html
>>
>> Kitware offers ITK Training Courses, for more information visit:
>> http://www.kitware.com/products/protraining.html
>>
>> Please keep messages on-topic and check the ITK FAQ at:
>> http://www.itk.org/Wiki/ITK_FAQ
>>
>> Follow this link to subscribe/unsubscribe:
>> http://www.itk.org/mailman/listinfo/insight-users
>>
>>
>>
>>
>> -- 
>> Juan Vidal Allende
>
> _____________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> Kitware offers ITK Training Courses, for more information visit:
> http://www.kitware.com/products/protraining.html
>
> Please keep messages on-topic and check the ITK FAQ at:
> http://www.itk.org/Wiki/ITK_FAQ
>
> Follow this link to subscribe/unsubscribe:
> http://www.itk.org/mailman/listinfo/insight-users

-- 
Gaëtan Lehmann
Biologie du Développement et de la Reproduction
INRA de Jouy-en-Josas (France)
tel: +33 1 34 65 29 66    fax: 01 34 65 29 09
http://voxel.jouy.inra.fr  http://www.itk.org
http://www.mandriva.org  http://www.bepo.fr

-------------- next part --------------
A non-text attachment was scrubbed...
Name: PGP.sig
Type: application/pgp-signature
Size: 203 bytes
Desc: Ceci est une signature ?lectronique PGP
URL: <http://www.itk.org/pipermail/insight-users/attachments/20100927/ff9f44c2/attachment.pgp>