ReTV at EBU MDN Workshop 2020
0 likes740 views
Presentation on "Metadata Driven TV Content re-purposing and re-publication" by Lyndon Nixon, MODUL Technology. Presented at the EBU MDN Workshop 2020
ReTV at EBU MDN Workshop 2020
- 1. Metadata Driven TV Content Re-Purposing and Re-Publication Lyndon Nixon, MODUL Technology EBU MDN 2020, 9/6/2020
- 2. Viewing of linear broadcastTV is decreasing while time spent with digital content on CatchupTV, on-demand OTT or social media rises. Broadcaster audiences are fragmented across digital channels and digital channels are full of competing content offers for their limited attention. TheTV industry is still catching up with their online competition in the use ofWeb technology: user tracking, personalisation and targeting.
- 3. ReTV develops aTrans-Vector Platform (TVP) to analyse content across all channels and publish to all media vectors with the effort of one
- 4. End UserTools Enabled ByTheTVP
- 5. Topics Compass: what are the trending topics to publish about?
- 6. Topics Compass metadata: keywords and entities International news articles Online pages aboutTV content on broadcasters Websites Social media accounts aboutTV content TV Program information from EPG data Related content (mentioning aTV program) from social media
- 7. Success metrics: frequency, impact, sentiment and WYSDOM
- 8. Prediction: what is the best topic to choose on a future date? Our events and anniversaries API highlights important events and anniversaries on a specific date. TheTopics Compass can identify time references in documents and aggregate those documents that refer to a specific date
- 9. Prediction: what is the best date to choose for a future topic?
- 10. Topics Compass 5min demo
- 11. Content Wizard: what content do I publish to get the audiences attention?
- 12. Content Wizard: re-purposing content for the channel
- 13. Video summarization 13 1. Video length restrictions (e.g. social media) 2. According to topic(s) (predicted to optimise success) 3. Guided by purpose (e.g. trailer to promote future content, highlights of past content) The SUM-GAN model Idea: learn keyframe selection by minimizing the distance between the deep feature representations of the original video and a reconstructed version Problem: how to define a good distance? Solution: train a discriminator network (GAN)! Goal: train Summarizer to maximally confuse the discriminator when distinguishing the original from the reconstructed video.
- 14. Content Wizard: recommending when to schedule the publication
- 15. Content Wizard: 5min demo
- 16. 4u2: how viewers can access new personalised services with your content
- 17. 4u2: how viewers can access new personalised services with your content
- 18. 4u2: 5min demo
- 19. The ReTV Stakeholder Forum is your opportunity to engage with us, be first to get updates and have the opportunity to test our tools and applications! I am here all day for live demos of any of the ReTV tools just send me a message in Skype (lyndonjbnixon) / e-mail (lyndon.nixon@modul.ac.at) Sign up for the ReTV Newsletter and get an update every few months from us!
- 20. @ReTV_EU Facebook: ReTVeuwww.ReTV-Project.eu Instagram: retv_project Dr. Lyndon Nixon ReTV Project Coordinator info@retv-project.eu @ReTV_EU @ReTVeu ReTV Project retv_project