Artificial intelligence is revolutionising content production

Although videos have become an increasingly important part of digital communications, their lifecycle is often short. Using technologies developed in the MeMAD project, video events can be converted into text and metadata on an industrial scale. Lingsoft's contribution to the project  focuses especially on speech recognition and indexing.

The lifecycle of audiovisual content is often short. Finding specific data on audiovisual materials is a slow process, and the manual improvement of searchability with subtitling and metadata, such as creating keywords and timing, are time-consuming and expensive. 

The three-year, EU-funded MeMAD project specifically focuses on the challenge media producers face in managing large volumes of video data. Using technologies developed in the MeMAD project, video events can be converted into text and metadata on an industrial scale. MeMAD improves the availability of audiovisual data for both content producers and viewers: subtitled videos serve the hearing impaired, language learners and anyone wanting to watch videos without sound − something we all want to do every now and then.
The automatic description of content in text can revolutionise the processing of large video collections and make it possible to recycle and reuse existing materials. “The new discoverability, accessibility and personal service experience made possible by artificial intelligence will be crucial success factors for the European media sector. Understanding content is one of the more promising areas in artificial intelligence right now," explains the Finnish Broadcasting Company YLE's Head of Innovation Strategy Anssi Komulainen.

Lingsoft is focusing specifically on the further development of artificial intelligence based on Finnish and Swedish speech recognition, speaker identification and the indexing of materials. In addition to Lingsoft and YLE, Aalto University, the University of Helsinki, the University of Surrey, INA, EURECOM and Limecraft are also participating in the project. MeMAD (Methods for Managing Audiovisual Data) is funded by the European Commission Horizon 2020 Programme. 

Follow the project on the website

The European Commission is the executive arm of the European Union. It promotes the general interests of the EU by, for example, participating in the drafting of EU strategy. One of the EU's strategic goals is to ensure its competitiveness in the global market, generate European growth and create new jobs. Every effort is being made to tackle this challenge by supporting science, the development and implementation of new technologies and innovations, and finding solutions to the biggest social challenges facing Europe. Promoting the achievement of this goal, the Horizon 2020 programme is the largest EU funding programme of all time, totalling nearly EUR 80 billion from 2014 to 2020.