W3C home > Mailing lists > Public > public-media-annotation@w3.org > March 2010

Action-221 - Check how many standards support content location

From: <Bennett.Marks@nokia.com>
Date: Mon, 15 Mar 2010 22:52:39 +0100
To: <joakim.soderberg@ericsson.com>, <werner.bailer@joanneum.at>, <vmalaise@few.vu.nl>, <wslee@etri.re.kr>
CC: <public-media-annotation@w3.org>
Message-ID: <15443028349C3C44BA819EB51B3B4F3C57264EAE92@NOK-EUMSG-01.mgdnok.nokia.com>
The following is a review of standards groups and how they deal with location. Of special interest is what level of detail is used when defining the semantics. This is to fulfill Action-221

Background

The definition of Location has not been semantically clear in the marketplace. There has been a large confusion between the concepts of Location Created and the location that is shown in the media. As an example, videos and still images geo-located to the Eiffel Tower are equally split between panoramas of Paris taken from atop the Eiffel Tower, to shots of the Eiffel Tower taken from all over Paris.

Workflow considerations, provenance from other standards bodies

Location Created is now routinely collected as an automatic contextual property via the recording of the latitude and longitude of the recording device.  In the case of video, this clearly leads to the issue of time-domain based recording, and how to best label a video clip with a single location.
The recording of location metadata and the semantic definitions used are highly dependent upon the contextual needs of the defining organizations.
Simple real world workflows, such as those defined by device manufacturers, GIS related activities, or those trying to document actual places and events tend to create the simplest and clearest (albeit most limited) definitions of location.

Exif – CIPA (Camera Industry Producers Association) – standard for the collection of still image technical and context information. Exif GPS IFD defines separate GPS locations for Location and Destination.

IPTC – International Press Telecommunications Council – defines the standard for still images and videos to be used in the news business. IPTC defines separate properties for location created and location shown, and defines a five level hierarchy for textual location. They cannot deal with fictional locations. 
  http://www.iptc.org/std/photometadata/specification/IPTC-PhotoMetadata(200907)_1.pdf 

Tools for converting from the IPTC five level hierarchy to and from latitude/longitude exist, such as    http://www.geoiptc.com/EN/Index.html


More complete, but still real world definitions of geo-location are defined by:

OGC – definer of GML ( Geographic Markup Language).  

Google – definer of KML (Keyhole Markup Language). KML has been submitted to the OGC 
Security issues have been explicitly linked to the capture and transmission of location data. This has resulted in work which explicitly deals with the policy and mechanisms around geospatial data.

IETF – Geopriv – deal only with geographic values, and define the geo: URI scheme for transmitting locations.

OASIS – owner of GeoXACML – a geospatial extension to XACML. This is related to the requirements from the Liberty Alliance. Extensions are supported in OGC GML 3.

Note: The relationship between the various standards is best described at http://www.opengeospatial.org/standards/is


In addition to the security and privacy considerations business considerations have also influenced the work defining the collecting, and transmitting of location data.

OMTP – The Open Mobile Terminal Platform – has described a telecom centric view of the requirements for creating and using geolocation data in a telephony environment.
http://internal.omtp.org/Lists/ReqPublications/Attachments/26/OMTP%20Positioning%20Enablers%20v1_0.pdf  . These requirements were meant as input to OMA work.
OMA – definer of SUPL - an over the air way to securely transmit location information.

Organizations that have specifically been dealing with location metadata as a property of media include:

MWG – Metadata Working Group – defining a simple set of metadata properties for consumers. They are separating out the definition of location created from location shown.

SMPTE – manages registry of metadata for professional media – SMPTE Metadata Dictionary. This dictionary defines location in many dimensions, and includes provisions for dealing with fictional locations.

MPEG - MPEG-7, MPEG-21 – General framework for creating metadata. Specific location information is viewed as an implementation detail.

BBC – SMEF-DM – Standard Media Exchange Format – Broadcast metadata model for managing broadcast assets. It consists of a data dictionary and set of Entity Relationship Diagrams. https://www.bbc.co.uk/guidelines/smef_dm_licence.shtml 

Variants and Qualifications
One of the reasons why Location Shown can be a problematic property is that videos especially can be of fictional locations. It is extremely difficult, if not impossible to create a vocabulary which can express this realm. And it should be noted, that in this case, expressing a fictional location using simple keywords does not represent a loss of fidelity.

   -Bennett


-----Original Message-----
From: ext Joakim Söderberg [mailto:joakim.soderberg@ericsson.com] 
Sent: Monday, March 15, 2010 5:25 PM
To: Marks Bennett (Nokia-CIC/Boston); werner.bailer@joanneum.at; vmalaise@few.vu.nl; wslee@etri.re.kr
Cc: public-media-annotation@w3.org
Subject: RE: Updated ontology doc within Seoul F2F meeting

Hello,
We want to keep the number of properties to a minimum.
Can we solve it by having ma:location and adding sub-types for Shown / Created ?

Regards
/Joakim

-----Original Message-----
From: public-media-annotation-request@w3.org [mailto:public-media-annotation-request@w3.org] On Behalf Of Bennett.Marks@nokia.com
Sent: den 26 februari 2010 18:11
To: werner.bailer@joanneum.at; vmalaise@few.vu.nl; wslee@etri.re.kr
Cc: public-media-annotation@w3.org
Subject: RE: Updated ontology doc within Seoul F2F meeting

In the still image world, the IPTC notion of "LocationCreated" and "LocationShown" is finally becoming ubiquitous. Indeed, even the Exif GPS properties have 2 separate areas for collecting point of creation and subject location (destination). This notion translates fairly well to video. The real issue is the widely variable syntactics assigned to location ( e.g. hierarchical data structures, country codes and controlled vocabularies). This is a result of the desire to provide extra semantic value to the location property, above what comes from treating the location as a keyword.

The MWG has found it extremely useful to clearly separate these two notions of location, and we hope MAWG will also make the semantic difference clear. Furthermore, I would hope that you review the decision to put depicted location into keywords, as there is a loss of semantic value when you do that.

-Bennett Marks  Sr. Architect CDO/CIC NOKIA Chair Metadata Working Group - Video › 5 Wayside Rd., Burlington MA  01803 š bennett.marks@nokia.com È +1 781 308 6556 [mobile]  +1 781 993 1911 [fax]
Skype: bennettmarks  Yahoo:bennettmarks439 GMail:bennettmarks439

-----Original Message-----
From: public-media-annotation-request@w3.org [mailto:public-media-annotation-request@w3.org] On Behalf Of ext Bailer, Werner
Sent: Friday, February 26, 2010 5:00 AM
To: Veronique Malaise; 이원석
Cc: public-media-annotation@w3.org
Subject: AW: Updated ontology doc within Seoul F2F meeting

Dear Véronique, all,

I agree that both describing recording and depicted location is useful. We had the discussion at the F2F in Stockholm, and we came up with the proposal to use location for recording location, and would put information about depicted locations into description or keywords (similar to other content related annotation).

Best regards,
Werner

________________________________________
Von: Veronique Malaise [vmalaise@few.vu.nl]
Gesendet: Freitag, 26. Februar 2010 10:18
An: 이원석
Cc: Bailer, Werner; public-media-annotation@w3.org
Betreff: Re: Updated ontology doc within Seoul F2F meeting

Hi everyone,

Well, one problem with this restricted definition is, for example that all the "spaghetti westerns" (a fistfull of dollars etc directed by Sergio Leone) would have the location "Italy" although the story is supposed to take place in Texas/Mexico. I would be in favor of keeping the extended definition: a resource can be about some place and be shot in another one. It is actually quite common for movies!

Best,
Véronique

On Feb 26, 2010, at 7:35 AM, 이원석 wrote:

Hi. Werner and all,
Thanks for your additional comment.

I agree with your comment. I will use this instead of earlier one ☺

Best regards,
Wonsuk.

From: Bailer, Werner [mailto:werner.bailer@joanneum.at]
Sent: Thursday, February 25, 2010 12:12 AM
To: 이원석; public-media-annotation@w3.org<mailto:public-media-annotation@w3.org>
Subject: RE: Updated ontology doc within Seoul F2F meeting

One more comment: The definition of ma:location is “A location associated with the resource. This property can refer to a depicted location or the location where the resource was captured.” We had changed that to “A location where the resource has been shot/recorded.” at the 4th F2F (see summary table). I prefer the letter, as it is a stricter definition.

Best regards,
Werner

From: public-media-annotation-request@w3.org<mailto:public-media-annotation-request@w3.org> [mailto:public-media-annotation-request@w3.org] On Behalf Of ???
Sent: Mittwoch, 24. Februar 2010 09:36
To: public-media-annotation@w3.org<mailto:public-media-annotation@w3.org>
Subject: Updated ontology doc within Seoul F2F meeting

below is the updated ontology doc within Seoul F2F meeting.
Please review and if you have any comment, let me know.

http://dev.w3.org/2008/video/mediaann/mediaont-1.0/mediaont-1.0.html


best regards,
Wonsuk.


Received on Monday, 15 March 2010 21:54:38 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 15 March 2010 21:54:39 GMT