Research indexes 70,000 reusable videos

With Videorooter we aimed at creating video fingerprinting technology: a means to recognise a video, regardless of changes in file format or size. With this technology, we created a white list of 70,000 Creative Commons licensed works and a white paper describing how we can use this technology to promote an open Internet.


8 December 2016

This year Kennisland and our partner Commons Machinery have worked on project Videorooter. The project, funded by the Dutch SIDN Fonds, aimed at creating video fingerprinting technology. A means to recognise a video, regardless of changes in file format or size. With this technology, we created a white list of 70,000 Creative Commons licensed works and a white paper describing how we can use this technology to promote an open Internet.

screen shot videorooter

screen shot videorooter

Maker: Kennisland

Rights:

Download
This hash (f803e30307270717fc040c1f0c2f3705fc040c1f0d2d3d050000307c7c300000) shows that a video is both hosted on Wikimedia Commons and the Internet Archive.

The project++WebsiteExplore the details of our work on videorooter.eu. creates the technological conditions to fingerprint video. These fingerprints allow you to create a database of provenance information: a copyright registry. Information about the origin, rights status and licensing options of the work. This information is needed when you want to contact the creator, want to do proper attribution, or license a work.

The Internet is a place of sharing and copying. Provenance information tends to get lost during this process, which usually does not happen intentionally. There have been some techniques that try to retain the link back to this information, maintaining a chain of attribution. These techniques include embedded metadata and watermarking. However, these techniques mostly do not work for any reproduction that alters the media file. For example, uploading to YouTube strips embedded metadata, and changes to the size of the file will often destroy a watermark.

WeFingerprinting allows you to create databases of provenance information. proposed and implemented a technique we coined as a persistent reproducible identifier. Such an identifier, a string of letters and numbers, can be created from a media file in the way it appears to the human eye. Humans can recognise images as the same regardless of size or format differences. Machines only see ones and zeros unless directed otherwise. The same image as a PNG and a JPG are completely different to a machine, but look the same to a human. The technology analyses video files like a human would see the file and can therefore create an identifier independent of file format and size.

This means you can create a dataset with fingerprints and match this with whatever information you want, like provenance information. We did just that++VideorooterSee also our project page.: we sourced over 70,000 videosExplore over 70,000 freely reusable videos on Videorooter. from Europeana.eu, Wikimedia Commons and the Internet Archive. On our site you can explore these works, upload your own videos to check if they appear in our database, and freely build upon our work.

Questions? Mail me at mz@kl.nl.

 

Deze tekst heeft een Creative Commons Naamsvermelding-licentie (CC BY) en is gekopieerd van de Kennisland-website. Ga voor de volledige versie met afbeeldingen, streamers en noten naar https://www.kl.nl/en/news/report-out-of-office-15-eu-copyright-reform-where-are-we-now/

This text has a Creative Commons Attribution License (CC BY) and has been copied from the Kennisland website. For a full version with images, streamers and notes go to https://www.kl.nl/en/news/report-out-of-office-15-eu-copyright-reform-where-are-we-now/