Provenance information in the blockchain

Posted by on Jun 7, 2016 in Uncategorised

Technologies for conveying provenance information are often conflated with blockchain technologies. As the blockchain is continuing to hype. Videorooter does not focus on the blockchain. The problem Videorooter seeks to address is a bit smaller, even if smaller doesn’t mean easy.

Doug Weldron: Blocks. CC BY-SA 2.0

Doug Weldron: Blocks. CC BY-SA 2.0

Videorooter works on provenance information attached to media on the Internet. Our focus is video, which has some specific challenges compared to other media, but the general idea is the same. Provenance information often gets lost when media is moved from platform to platform or when media is copied and used in new productions. Copying without provenance information often happens without the intent of the copier, but media is sometimes also copied without appropriate compensation (in terms of attribution, remuneration, or otherwise) to the original creator, for purposes that benefit the new user.

With the rise of blockchain technologies we see the promise of registering these works in a public ledger. The public ledger addresses issues of ownership and provenance. Blockchain technology forms a decentralised ownership system and allows transparency in transactions and is therefore attractive for a public ledger of provenance.

Blockchain technology conveys trust in transactions, like the transactions of a good, licensing of an artwork or moving of cryptographic money such as bitcoin. In a blockchain these transactions are self-contained. That is to say that all transactions refer to social constructions within the system. Whether it is money or intellectual property. As soon as a blockchain refers to social constructions outside the blockchain, the blockchain can fail.

Let’s consider an example: a photographer enters a photograph in the blockchain public ledger. This creates a public record of ownership of that photograph. The public ledger keeps track of any transaction based on the public record of ownership. However having a public ledger of transparent transactions of ownership and permissions is only one of the problems the issue of provenance and ownership on the Internet. The problem relating to provenance that Videorooter is addressing is identifiable media files. Media can be stored in dozens of file types and sizes. An image can be perceptually the same without being bit by bit the same file. For example when you convert an image from jpg to png, or resize a video. Each of these operations generate a new digital file. Without being able to link a file to an entry in a public ledger, provenance, ownership and permission information will continued to be lost during copying and republication of media on the Internet.

To be able to identify digital works that are the same in terms of human perception, we need mechanisms, like (but not limited to) perceptual fingerprints. While blockchain is rapidly maturing, the field of perceptual fingerprinting is not. And if it is maturing it is not shared publicly (see our article on Content ID). This creates a dangerous weak link in the blockchain. Videorooter wants to create the conditions where we can share our perceptual fingerprints across projects and so help to create the opportunity for new projects and new services. These fingerprint strengthen the – blockchain based – public ledgers for transparently storing provenance, ownership and permission information about media .

Fingerprinting technologies do not replace blockchain technologies and we’re happy others are working on the problems inherent to other parts of the provenance topic. The problem Videorooter seeks to address is quite a bit simpler, even if simple doesn’t mean easy: finding ways which enable all of the different initiatives to share and exchange fingerprints about (video) files.

Do you want to help us create strong links from the blockchain to external files? Take a look at our workshop in September or contact me at mz@kl.nl.

Leave a Reply