LinkedTV at ACM #TVX2014 showing Linked Television as a new form of TV and online video interactive experience

LinkedTV, as a leading EU research project in the area of TV-Web convergence and future forms of interactive television,
is pleased to be present at this weeks ACM TVX 2014 conference with 1 conference paper, 1 work in progress poster, 1 workshop demo and 1 conference demo.

ACM TVX 2014, formerly EuroITV, is the key conference in this area, bringing together researchers and industry practitioners to
present, discuss and explore new forms of interactive experience for television and online video.

It all starts Wednesday at the Workshop on Interactive Content Consumption (WSICC), which LinkedTV’s workshop series on
Future Television was merged into this year, where our Media Arts scenario represented by University of Mons is
giving a multimodal, interactive demo on the “Social Documentary: An interactive and evolutive installation to
explore crowd-source media content”.

Then on Thursday in the main conference session on “Consumption: Multiple Screens and Attention” joint work between
LinkedTV partner University of Mons and an external partner will present results on “Visual Attention Measures for
Multi-Screen TV”. This is an important contribution on visual attention for multi-screen TV applications, such as
could be built using LinkedTV’s own Multi-Screen Toolkit.

Later in the Work in Progress poster session, scientific coordinator Dr Lyndon Nixon (MODUL University) will be
present with the LinkedTV poster and showing LinkedTV, seamlessly integrating TV programming with related Web
content. It just gets better, since after that session and during the drinks reception, LinkedTV will be giving one
of the accepted demos, with Lilia Perez Romero (CWI) and Dr Nixon to hand to demonstrate linked television with
news (RBB) content.

LinkedTV looks forward to meeting interested researchers and practitioners in Newcastle at the very first ACM TVX 2014!

Posted in Demos News Presentation by No Comments

Orwellian Video Recommendation: LinkedTV partner demo at #PAIS2014 #ECAI2014

LinkedTV is pleased to announce that project partner UEP (University of Economics Prague) has had a system demo accepted at the Conference on Prestigious Applications of Intelligent Systems – PAIS 2014 taking place on 20–21 August 2014, Prague, Czech Republic co-located with the 21st European Conference on Artificial Intelligence (ECAI 2014) with the intriguing title of

“Kinect Eye: Orwellian Video Recommendation”

A public screencast is available!


Posted in Demos Events News by No Comments

Social Events in Web Multimedia workshop @ICMR2014 proceedings published

The Proceedings of the 1st International Workshop on Social Events in Web Multimedia @ ACM International Conference on Multimedia Retrieval (ICMR 2014) (Glasgow, UK, April 1-4, 2014), which was co-organized by EU projects MediaMixer, LinkedTV and SocialSensor, are now publicly accessible from the workshop’s web site (direct link to the proceedings: ). 

Posted in Uncategorized by No Comments

LinkedTV with FOUR papers at #LREC2014 reflects relevance of this domain to Linked Television

The LinkedTV project is pleased to be present at the 9th edition of the Language Resources and Evaluation Conference (LREC 2014, 26-31 May, Reykjavik, Iceland) with no less than FOUR accepted papers! We warmly congratulate our LinkedTV partners and wish them the very best for their time at the conference! This reflects the cutting edge research on topics relevant to that domain which also form part of the innovative solution being created by partners for “Linked Television”, such as improving Automatic Speech Recognition (ASR), keyword detection from text and Named Entity Recognition (NER). 

  • T. Kliegr., O. Zamazal Towards Linked Hypernyms Dataset 2.0: complementing DBpedia with hypernym discovery
  • M. Stadtschnitzer, J. Schwenninger, D. Stein and J. Köhler. Exploiting the large-scale German Broadcast Corpus to boost the Fraunhofer IAIS Speech Recognition System
  • Giuseppe Rizzo, Marieke van Erp and Raphaël Troncy. Benchmarking the Extraction and Disambiguation of Named Entities on the Semantic Web
  • T. Kliegr, V. Zeman, M. Dojchinovski. Linked Hypernyms Dataset – Generation Framework and Use Cases. (In the Linguistic Linked Data Challenge)
Posted in Uncategorized by No Comments

All slides from our #lime2014 workshop about #LinkedMedia at #eswc2014 now online

The 2nd Linked Media workshop gathered 20 participants – possibly the best attended workshop at ESWC 2014 – for a great set of presentations and good discussions in-between and at the end. With thanks to all presenters, we have now embedded all workshop presentation slides on the LIME 2014 webpage.


Posted in Uncategorized by No Comments

Survey of Online Tools for Semantic Media Annotation – LinkedTV leads the way

Presented today by LinkedTV scientific co-ordinator Lyndon Nixon at the 2nd Linked Media workshop during ESWC2014, the Survey of Semantic Media Annotation Tools considered current tools and services available to those interested in generating interoperable rich (semantic) descriptions of their media at the fragment level. LinkedTV has published the slides to SlideShare:

The Linked Media principles consider consensus on the approach to metadata for online media as vital if access to and re-use of Web based media is to prove as successful as the text based Web, where HTML enabled Web browsers to have a common understanding of descriptions of Web pages. Those principles cover

Since LinkedTV work on media annotation and linking is driven by the same partners who are promoting the vision of Linked Media, it is not surprising that it is LinkedTV outputs such as NERD, TV2RDF and the Editor Tool which are closest to those principles and can be part of a technology stack which can enable media owners to generate Web-ready media annotations which can be used by applications for media retrieval, browsing and enrichment.


Tags: ,
Posted in General Presentation Uncategorized Vision by No Comments

#eswc2014 Accepted Posters and Demos published: FIVE LinkedTV submissions will be shown!

The full list of Posters and Demos to be shown next Thursday at ESWC2014 has been published, and LinkedTV is proud to note that no less than FIVE submissions from the project will be shown there!

We present an approach that leverages on the knowledge present on the Web for identifying and enriching relevant items inside a News video and displaying them in a timely and user friendly fashion. This second screen prototype (i) collects and o ffers information about persons, locations, organizations and concepts occurring in the newscast, and (ii) combines them for enriching the underlying story along fi ve main dimensions: expert’s opinions, timeline, in depth, in other sources, and geo-localized comments from other viewers. Starting from preliminary insights coming from the named entities spotted on the subtitles, we expand this initial context to a broader event representation by relying in the knowledge of other Web documents talking about the same fact. An online demo of the proposed solution is available at

Entities play a key role in knowledge bases in general and in the Web of Data in particular. Entities are generally described with a lot of properties, this is the case for DBpedia. It is, however, difficult to assess which ones are more important than others for particular tasks such as visualizing the key facts of an entity or fi ltering out the ones which will yield better instance matching. In this paper, we perform a reverse engineering of the Google Knowledge graph panel to find out what are the most important properties for an entity according to Google. We compare these results with a survey we conducted on 152 users. We fi nally show how we can represent and explicit this knowledge using the Fresnel vocabulary.

In this paper we present LiFR, a lightweight DL reasoner capable of performing in resource-constrained devices, that implements a fuzzy extension of Description Logic Programs. Preliminary evaluation against two existing fuzzy DL reasoners and within a real-world use case has shown promising results.

This paper introduces the LinkedTV User Model Ontology (LUMO), developed within the LinkedTV EU project. LUMO aims to semantically represent user-pertinent information in the networked media domain and to enable personalisation and contextualisation of concepts and content via semantic reasoning. The design principles of LUMO and its connection to relevant ontologies and known vocabularies is described.

Increasingly, European citizens consume television content together with devices connected to the Internet where they can look up related information. In parallel, growing amounts of Linked Open Data are being published on the Web, including rich metadata about its cultural heritage. Linked Data and semantic technologies could enable broadcasters to achieve added value for their content at low cost through the re-use of existing and extracted metadata. We present on-going work in the LinkedTV project, whose goal is to achieve seamless interlinking between TV and Web content on the basis of semantic annotations: two scenarios validated by user trials – Linked News and the Hyperlinked Documentary – and a companion screen application which provides related information for those programs during viewing.

Join the session Thursday morning at ESWC, and if you can’t be there, follow Twitter @linkedtv and check news at for updates from the conference!


Posted in Uncategorized by No Comments

Linked Television is present at #eswc2014 Workshops, Posters, Demos, Project Networking and Sponsorship!

Linked Television will be difficult to avoid this year at the Extended Semantic Web Conference 2014, taking place from May 25 to 29 in Crete, Greece. LinkedTV partners and work will be present in the workshops, posters and demos sessions, the LinkedTV project will participate in the Project Networking and a LinkedTV demo will be accessible during the main conference at a sponsor’s table. We look forward to showing a research and industry audience from the Semantic Web domain how we apply semantic technologies in an unique area: television!

In detail:

  • on Sunday, May 25 the 2nd Linked Media workshop will take place, with a keynote from Sebastian Schaffert on linking Media and Data with Apache Marmotta, and then a mixed program of presentations and demos related to media annotation and Linked Data;
  • on Tuesday, May 27 the Project Networking session will build synergies between 30 European funded projects and all, including LinkedTV, will present themselves in the evening;
  • on Thursday, May 29 the Poster and Demo session will include many aspects of LinkedTV research and development (see the list below);
  • throughout the main conference, as an ESWC sponsor, LinkedTV will demo its Linked Television experience (powered by semantics!) at a sponsor’s table (TBC, possibly in the registration area)

LinkedTV posters and demonstrators at ESWC2014 (updated):

  • A companion screen application for TV broadcasts annotated with Linked Open Data Nixon L., Belice Baltussen L., and Perez Romero L. and Hardman L.
  • Augmenting TV Newscasts via Entity Expansion Redondo García J.L., Hildebrand M., Perez Romero L., Troncy R.
  • What are the Important Properties of an Entity? Comparing Users and Knowledge Graph Point of View Assaf A., Atemezing G., Troncy R. and Cabrio E.
  • LUMO: The LinkedTV User Model Ontology Dorothea Tsatsou and Vasileios Mezaris
  • LiFR: A Lightweight Fuzzy DL Reasoner Dorothea Tsatsou, Stamatia Dasiopoulou, Ioannis Kompatsiaris and Vasileios Mezaris

We look forward to meeting you there!

Tags: , , ,
Posted in Demos Events General Presentation by 1 Comment

CERTH releases video shot segmentation software

CERTH just publicly released an advanced edition of its software for the automatic temporal segmentation of videos into shots, i.e., the division of the video in sequences of consecutive frames captured without interruption by a single camera. The released software can detect both abrupt and gradual shot transitions with high accuracy, and makes use of CUDA-based parallel computing capabilities of modern PCs for making the whole analysis process at least 2-3 times faster than real-time processing (depending on the processing capability of the PC’s graphics card). The development and release of this software was supported by EU research projects LinkedTV ( and MediaMixer (

The software is available for download at

Related technical paper: E. Apostolidis, V. Mezaris, “Fast Shot Segmentation Combining Global and Local Visual Descriptors”, Proc. IEEE Int. Conf. on Acoustics, Speech and Signal Processing (ICASSP), Florence, Italy, May 2014. (available at

Related demos:

Tags: ,
Posted in Demos News by No Comments