Motivated by the European Union’s challenge to develop a technological system that would make it easier for European citizens to gain access to cultural resources CEPIC has developed the CEPIC Image Registry (CIR) that will allows simultaneous visual or text search across thousands of image databases.
A pilot of this registry was demonstrated at the CEPIC Congress in Barcelona. Using image fingerprinting technology, the registry makes it possible for potential image users who find an image they would like to use, either online or in print, to locate the copyright holder and pay for the use.
Once this registry is fully operational customers will no longer be able to claim that an image is an “orphan work” if the creator or a licensor can be found by using the CIR. As an added benefit users will be provided with the creator’s name, caption information and other metadata that is usually unavailable when images are found on the Internet.
It is important to note that the CIR
will not license rights to images or be used in any way as a sales tool. Potential users will be directed to the copyright holder, or the licensing organization that normally licenses work for the photographer.
Among the aims of this registry are: (1) to enable the retrieval of information of the copyright status, identification of appropriate rights holders and data about the works and (2) facilitate the clearance of rights for works and including out-of-distribution-works in view of their digitations and online accessibility
How It Works
The CIR is a central location (in some senses like Google Image search) where image requests can be submitted. The request will be delivered simultaneously to all the databases that have opted-in to the CIR. The big difference between CIR and Google is that Google searches the entire web, not selected databases. Google then delivers millions of images as they appear on web pages and makes no distinction as to whether or not a license is required to use the image. CIR will only deal with databases containing the work of creators that want to control and/or license the use of their images.
Once a stock agency decides that it wants to be included in the CIR there are two ways of doing it.
1. By joining an existing ‘Source’ (at present there are two Sources available ‘Album/THP’ and ‘ImageIRC/PicScout’). All the agency’s images will be ‘fingerprinted’ by the Image Search Engine of the selected Source. This is Album Reverse Image Search (Album RIS) or PicScout’s Image IRC software. As new images are added to the agency’s collection and therefore to the Source, they are also fingerprinted by the Source selected. The fingerprints are stored by AlbumRIS or ImageIRC.
2. By creating a new Source. To create a Source to be connected to the CIR the requirements are:
- To have an Image Database
- To have a Text Search Engine
- To have an Image Search Engine (similar to PicscoutIRC or AlbumRIS)
- To conform the CEPIC Communication Protocol
This service is supplied free of charge. Each image fingerprint is also assigned a unique 19 digit CUIN (CEPIC Unique Identification Number). The first five digits will identify the Source Search Engine making it possible to link up to 99,999 collections. The other 14 digits are used to specify the specific image within the collection. If the same image is available through multiple distributors the CUIN number will be different in each collection.
When a customer drags and drops an image on the CIR search tool that image is immediately fingerprinted and the fingerprint is automatically forwarded to all the Source search engines.
Each database compare the fingerprint with the fingerprints of its images to find the exact matches. It also finds versions of the specific image that have been flipped, colorized or are cropped sections of the image. It will not find near similars. All the thumbnails and all the metadata including the creator’s name that is associated with that image are returned to the CIR and displayed for the customer’s consideration.
All this happens in milliseconds. Click on any of the thumbnails and a preview will be opened. On the right side of the screen will be a listing of every agency where the image can be licensed. Each listing includes: the agency image code, caption, photographer’s name, licensing restrictions if any, the CUIN number and a link to the agency where a full size preview and licensing information can be found. The CIR will never license usages. It will simply provide customers with information about where to go to properly license a usage.
The agencies are listed in the order that they respond to the request. Those with greater bandwidth or hosted near the CIR server are likely to always be first. Currently the server is located in Germany. Many image suppliers currently have their images with many distributors. Thus, it is possible that in some cases 200 or more versions of a specific image might be returned, each with a unique CUIN number.
Stock agencies must Opt-In to be included in the CIR. The CIR does not include anyone that doesn’t want to be included.
While this service is designed to represent stock agencies, and the creators they represent, other image creators and owners may find the service helpful. There is no indication that any limits will be placed on who can join.
Funding
The initial plans are that there will be no charge to participate in the registry. The costs may be somewhat limited since the registry is not building a database of images. But despite this fact there will be costs as the registry grows. It is not clear where the funds will come from to cover these costs. CEPIC may fund the project initially as a member service, but it is not clear if non-members of the organization will be allowed to participate for free. Hopefully, a viable business plan will be developed in the near future.
Operational Control
The CIR is expected to be a non-profit organization, but that says nothing about who will have operational control. Some are concerned that board members who come from within the industry will be able to establish policies that favor their organizations over those of others. Some believe the operations should be turned over to an independent or government organization, but no one has explained exactly how this would work.
Price Comparison
Another concern is that by using this tool customers will be able to very easily compare prices from different distributors. This may tend to drive down the fees charged for images that are widely distributed. However, this does not seem to be a valid reason for trying to scuttle the project. Rather, if the CIR becomes widely used it may require some dramatic changes in the way distribution networks are set up compared to the way they operate today. Those who want to place some controls on the fees charged to use their images may want to be more careful about who they allow to license them.
I went on the test site and looked for image CUIN 2000000000000001993 which is a self-portrait of Rembrandt and is captioned “Rembrandt and Saskia in the parable of the Prodigal Son. Oil on canvas (1635-1639). 161 x 131 cm.” The following are the images that appeared on this pilot site that has only two source collections.
| Copyright Holder |
Licensing |
Pricing |
| |
Type |
|
| Getty Images |
RM |
$252 for Webisode, higher for other uses |
| Masterfile |
RM |
$250 minimum, higher depending on use |
| Jupiter |
RM |
$36 to $361 depending on file size |
| Thinkstock |
RF |
$49 for 5 image pack, Subscription $139 per month |
| Photos.com |
RF |
$7.99 for single use |
| Universal Image Group |
RM |
no easy access to pricing |
| THP |
RM |
no easy access to pricing |
| AGE Fotostock |
RM |
no easy access to pricing |
| Album |
RM |
no easy access to pricing |
| Superstock |
RM |
no easy link to the site |
Where do you think the customer will go to purchase this image? Note that Getty Images owns Jupiter, Thinkstock and Photos.com. Getty is offering this same image at anywhere between $7.99 for unlimited use to well over $252 depending on how it is used. Certainly no customer in their right mind would go to Getty Images to license this image. After a few searches, and looking at the other options offered, they may decide that they don’t need Getty Images for much at all. Of course if a customer who needs that images doesn’t know about CIR they may search the Getty site, find the image and license it because they are unaware of the alternatives.
If this service is widely adopted by buyers, sellers may not need to place their images with multiple distributors around the world. In fact it may be a disadvantage unless sellers are allowed to place some restrictions on distributor pricing.
Promotion
It is not clear at this stage how this site will be promoted. Some of the major publishers will learn of it easily, but there is a huge amount of unauthorized use by smaller players, who are both within the industry and on the fringes. Making them aware of this resource will be a challenge. Three-and-a-half years ago PACA launched PACASearch that now represents over 170 million images from over 60 archives.
PACASearch only offers keyword search and doesn’t include the very valuable image search that is a part of CIR. Nevertheless, if could be a very useful tool for finding images. However, it has had very little market penetration because it has lacked the funds to promote it and make customers aware that it exists. It seems likely that this is a problem CIR will face as well.