Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[protege-owl] Room ontology

James Howison james at howison.name
Mon Feb 2 09:16:14 PST 2009


I think it looks good; you are not using class restrictions etc, so  
it's really simple RDFS, which is good, you can get 'trickier' with  
OWL as you move forward.  Realistically this knowledge base is a small  
part of your overall project.

Your attachment mentioned slots. Keep in mind that RDF doesn't have  
slots.  You can't assume that because you declare the domain of a  
class that all members of that class will have that property, that  
only works the other way around.  Of course you can enforce that  
constraint in your own code in various ways.  See this paper:

http://clarkparsia.com/weblog/2008/07/28/integrity-constraints-in-pellet/

--J

On 2 Feb 2009, at 8:34 AM, babar shahzad wrote:

>
> sorry, I forget attachment, you can fine it here.
>
>
> Babar Shahzad Chaudary,
> +358408532112
> Deptt of Information Processing
> Science,
> University of Oulu, Finland.
>
>
>
> ----- Original Message ----
> From: James Howison <james at howison.name>
> To: User support for the Protege-OWL editor <protege-owl at lists.stanford.edu 
> >
> Sent: Monday, January 12, 2009 6:21:32 PM
> Subject: Re: [protege-owl] Room ontology
>
> Ah, well, that is easier.  You need a vocabulary for describing
> objects and their relations to each other.
>
> So you are looking for a vocabulary for describing furniture and
> moveable objects (such as balls etc).  Everything will have a
> position, using some x,y,z coordinate scheme.  Sounds like everything
> will need an "affordances" type set of statements.  For example a
> table, having a flat upper surface, can have a ball placed on top of
> it.  But a ball, not having a flat upper surface, couldn't have a
> table placed on top of it.  You might also need to describe structural
> strength (eg a blow-up chair cannot have a safe on top of it, but a
> safe could have a blow up chair on it).
>
> At some point you are describing a set of arbitrary complex polygons,
> as well as surface textures etc.
>
> I'd encourage you to simultaneously search for relevant ontologies
> (furniture, surfaces, angles, materials), while simultaneously
> developing your own prototype ontology, using Protege.
>
> eg (in Turtle syntax)
>
> :table1 rdf:type :Table ;
>         :affordance :flat_upper_surface ;
>         :height "1.2" ;
>         :width "1.2" ;
>         :length "1.2" ;
>         :position_x "34.2" ;
>         :position_y "33.3" ;
>         :position_z "0" .
>
> You could then have a restriction so that anything with a :position_z
> attribute of 0 is also rdf:type :OnFloorObject
>
> Come to think of it, fully three-d positioning is more complex than
> this, since you need to know the full shape.  You want the ability to
> describe polygons, and link them together into objects.  I imagine
> that your "augmented reality" system has a coordinate scheme that it
> is expecting, so I'd start there.
>
> Good luck :)
>
> --J
>
> On 12 Jan 2009, at 3:51 AM, babar shahzad wrote:
>
>>
>> Actually its like that we don't need to find the object/objects'
>> position on  the table. As we have to display virtual objects later,
>> So our ontology need to convey the information to augmented reality
>> application that there there, these these objects are placed on a
>> table.
>>
>> I think this information will make it easy for you to help me.
>> Thanks a lot in advance for any kind of help.
>>
>> Babar Shahzad Chaudary,
>> +358408532112
>> Deptt of Information Processing
>> Science,
>> University of Oulu, Finland.
>>
>>
>>
>>
>> ________________________________
>> From: James Howison <james at howison.name>
>> To: User support for the Protege-OWL editor <protege-owl at lists.stanford.edu
>>>
>> Sent: Friday, January 9, 2009 8:27:27 PM
>> Subject: Re: [protege-owl] Room ontology
>>
>> Perhaps it will be easier to help if you tell us more about your
>> intended algorithm?  I mean what do you know about the room
>> (dimensions, location) and the objects (location?).  One simple way
>> would be to use GIS type functions to detect if an object is in a
>> room.
>>
>> Or perhaps you know if an object is on a table, then you just need to
>> know whether the table is in the room ... see where I'm going with
>> this?
>>
>> --J
>>
>> On 9 Jan 2009, at 12:14 PM, babar shahzad wrote:
>>
>>> Hi,
>>>
>>> I am new to OWL & Protege.
>>> I need to build a room ontology such that ontology is able to detect
>>> 5-6 objects in a room. Like one table & 5-6 different objects placed
>>> over it of size of a tennis ball, like paper weight.
>>> Any help in building this ontology will be appreciable by me.
>>> Can any one refer some good source for it.
>>>
>>> I thank to you all in advance.
>>>
>>> Babar Shahzad Chaudary,
>>> +358408532112
>>> Deptt of Information Processing
>>> Science,
>>> University of Oulu, Finland.
>>>
>>>
>>>
>>> _______________________________________________
>>> protege-owl mailing list
>>> protege-owl at lists.stanford.edu
>>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>>>
>>> Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
>>
>> _______________________________________________
>> protege-owl mailing list
>> protege-owl at lists.stanford.edu
>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>>
>> Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
>>
>>
>>
>> _______________________________________________
>> protege-owl mailing list
>> protege-owl at lists.stanford.edu
>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>>
>> Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
>
> _______________________________________________
> protege-owl mailing list
> protege-owl at lists.stanford.edu
> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>
> Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
>
>
>
> <Shopping Scenario.JPG>_______________________________________________
> protege-owl mailing list
> protege-owl at lists.stanford.edu
> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>
> Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03




More information about the protege-owl mailing list