Radio Televisión Española (RTVE), the Spanish main public broadcaster, and VSN will present the advances of their joint project for automated metadata creation of content from the network’s archive through AI tools. This presentation will take place on Thursday 21st September at FIAT/IFTA World Conference.
Carmen Pérez, Deputy Director of Innovation and Technological Strategy, and Virginia Bazán Gil, Project Manager of the Documentary Archive, will represent RTVE. On the other hand, Antonio Brotons, VSN Project Manager, will show the processes that the company is deploying. During the FIAT/IFTA World Conference, media archive professionals and experts from all over the world join together to discuss the latest trends and innovations in media asset management, digital preservation, and media content access.
During the presentation, RTVE and VSN will illustrate the different phases of the project, its workflows, and how this technology allows an increase of efficiency in the catalog of RTVE’s archive. With that goal in mind, RTVE put out to public tender a project to apply automated metadata using Artificial Intelligence to 11,000 hours of content from its Archive, including material produced by TVE in the 60s and 70s.
Thanks to the integration with the main AI engines of the market and its adaptation to allow a cloud service, VSNExplorer MAM was the perfect system for the project. Therefore, VSN decided to present its candidacy for the tender along with the AI engine of Etiqmedia. During the bidding process, VSN obtained the best rating combining the technical and economic aspects and was finally awarded with the project.
Artificial Intelligence working for the metadata
The project has two phases: A testing process, which was developed during the last four months, and the final one-year project implementation starting in October. During the first one, RTVE and VSN defined the workflows between teams and completed the installation of the equipment. In addition, 160 hours of audio and video were successfully processed in this testing phase.
During the process, the files from RTVE Archive are ingested in VSNExplorer MAM. The system creates a previous asset with the media file and an XML file with information about the content provided by the network. Once the content is in the system, the AI engine starts generating the content’s metadata. When the process is over, all the information extracted is displayed in a single and centralized interface.
“From a technological point of view, this project was a big challenge for RTVE. For the first time, we deployed a system that allows the content management of our Archive using cloud solutions. Therefore, we needed the implication of different areas of the Corporation”, stated Carmen Pérez, RTVE Deputy Director of Innovation and Technological Strategy.
In terms of audio, this technology is capable of retrieving a complete text transcription in just a few minutes, its capitalization and accentuation, and also the recognition of people, places, events, products, organizations, and dates that are mentioned. Furthermore, it extracts an automated catalog of the content and the main keywords. On the video side, the AI allows facial recognition and identifies and catalog the scene, along with the objects, tags, and signs that appear in the footage.
All this information is available for quality control on the VSNExplorer MAM web-based interface. Therefore, users can review, check and edit the results to adjust them to the correct catalog parameters in a simple and fast way. For example, we can correct the audio transcription or introduce characters that were not recognized by the AI. Once this check is completed, the system automatically calculates the transcription WER (Word Error Rate) comparing the results obtained by the engine with the cataloger’s correction.
“For the RTVE Documentary Archive, this project does not only give access to thousands of hours of great content. It is also a unique opportunity to influence the development of technologies that will be common in the future. In addition, we are pioneers in the implementation of this technology in the daily workflow of the Archive”, said Virginia Bazán Gil, Virginia Bazán Gil, Project Manager of the RTVE Documentary Archive.
After all these steps are completed, VSNExplorer MAM creates a number of XML Files with all this information that goes back to the RTVE Archive. This way, the assets of the public network incorporate new metadata that vastly improves the features of the catalog, allowing for searches and retrieval of by RTVE users.
“RTVE is a reference of the Broadcast & Media industry for its constant search of innovative solutions. At VSN, we are delighted to have, once again, the trust of a leading organization like RTVE”, said Antonio Brotons, Project Manager of VSN. “This project shows all that the AI integrated into a MAM system can provide to a content catalog in large documentary archives”.
This project joins the list of the collaborations between VSN and RTVE. Recently, VSNExplorer MAM was also selected by the network for the Innovation HUB project. This initiative studies the use of cloud-based tools to create news content in small news production centers.
Subscribe to our newsletter to stay updated about our activity