ARGOS: Aiding insurers and law enforcement recover stolen art

    Formed by an association of French insurance companies, ARGOS keeps watch over valuable items placed up for sale in order to spot items which may be stolen property.   ARGOS confirms the effectiveness of LTU visual search in the field of stolen property investigations:   LTU's most advanced image matching has proven effective in stopping illegal traffic.

 

Paintings, furniture, family heirlooms ...
every year thousands of  valuables are stolen.

Many will disappear forever, sold somewhere in the world on the black market or perhaps placed up for sale with legitimate dealers, where they thus appear in printed catalogues or on dealer websites.

And in the end, large indemnities are paid to insured property owners, who would have much preferred to recover  the stolen object rather than receive the compensation.  Objects such as these may have enormous sentimental value.  

Not long ago, only a very rigorous investigation - with fastidious and costly research and time-consuming manual comparison of images - might result in recovering the object. 

But these investigations are now expedited with advanced technology.  ARGOS and LTU technologies have applied new methods of research and identification and have achieved results previously unattainable.  

 How to identify rapidly a stolen object
among items placed up for sale around the world:

     

identification by
image
recognition
LTU engine


 
  
  
     


 
 
ARGOS manages a database called FIBAR (Fichier Informatique des Biens Assurés Recherchés) which is made up of images and data concerning stolen objects.  This data is provided by the insurance company members of ARGOS.   The images in FIBAR are typical of images which are provided by property owners after a theft:  often the only  image available of an object is not ideal for an investigation:  the image may be of poor quality, such as an image of a painting in a living room, where the painting appears only as an incidental detail in the background.   Yet these images of variable quality must be compared with all of the images of objects which appear in the marketplace:  internet sites, dealer catalogues.  The must also be compared with compared with items surfacing in investigations by law enforcement organizations, such as the French agency OCBC (Office Central de lutte contre le trafic des Biens Culturels).


Investigations have these typical constraints :

TIME:   the sooner an object is identified

  • lesser the risk that the object will disappear into the various sales channels, illicit or legitimate
  • greater will be the number of recoveries and restitutions.
DATA:  the quality of data is the essential factor in these investigations.
 


 
"Insurers have known for a long time that the IMAGE is a particularly powerful element for investigations - for the simple reason that with an image there are no language barriers, " say Jean-Louis Marsaud, Director General of ARGOS.   "But the explosion in the number of images on the Internet as well as the accelerated activities of criminals, and they also use new technologies, these became the major challenge for ARGOS. "
     

 

     

In a few seconds, LTU engine
analyzes hundreds of thousands of images …

In 2006, ARGOS had decided to make the choice for new technologies, and it selected LTU technologies and its platform for

image recognition &
visual content search: LTU engine.

  
     

This advanced technology is based on a very simple principle!  just like the human eye, the engine analyzes the content of an image (a "query" image, captured on the Internet or from a catalog by ARGOS.  The engine then creates a descriptor of the image, called the DNA of the image.  This DNA (which contains information on the forms in the image) them may be compared with the reference database (the objects declared stolen).  Whatever be the quality of the image furnished, LTU engine will provide search results arranged by degree of similarity and determine if there is a match between identify objects.


The results:   " In hardly a few seconds, LTU engine analyses the 16,000 images furnished by our members and compares them every day with the hundreds of thousands of images which we capture on the Internet or receive in catalogues or from investigations. "  explains Mr. Marsaud.  "One of the major advantages of LTU engine is that we can consider increasing the size of the FIBAR database as well as the number of catalogues to analyze, without needing additional budget or additional human ressources. "		  
"Today, on matches of two-dimensional objects, such as pictorial objects (paintings, drawings), we observe a clear improvement in the quality of searches effected by ARGOS.  But it is clear that the LTU tool will be become even more effective as the FIBAR database will be largely expanded.  The purpose of ARGOS is the search and identification of stolen objects, so that they might be returned to their rightful owners.   Thanks to LTU technologies and to the willingness of insurers to equip themselves with the moste effective technologies, we succeed in helping insurers provide a better service to their clients.   And this also contributes to fighting crime and illegal traffic, concludes Jean-Louis Marsaud. 

About ARGOS

ARGOS is a not-for-profit organization created by French insurance companies.  Its objective is the search, identification, and recovery of vehicles and other valuables declared stolen for insurance claims.  ARGOS works exclusively for the benefit of its member insurance companies and in close cooperation with public authorities, notably the French governement agency OCBC  (Office Central de lutte contre  le trafic des Biens Culturels) another user of LTU technologies.

For more information : www.gieargos.org

 

 

DNA Art Image licensed with permission from DNA 11 Inc - www.dna11.com