National Association of Broadcasters

winner of the NAB Technology Innovation Award 2020

Workplan and Deliverables

Work Package 1 – Management
Lead: Fincons

Work package 1 addresses the overall coordination, management of the financial, administrative, technological, and operational aspects of the project, innovations management, as well as quality monitoring of the project’s results. It covers all the management activities and provides and maintains the link to the EC offices.

Work Package 2  – Specification and requirements development
Lead: HFC

The specifications and requirements for the CONTENT4All system architecture, the reference scenarios, collected data and formats are defined in work package 2. These requirements will be utilized as a starting point for the system modules that will be developed in work package 3 and 4, and for planning of work package 5 use cases, development work and integration efforts.

Deliverables available:

D2.1 Initial reference system architecture

D2.2 Initial requirements and specifications for pilot demonstrators

D2.3 Final reference system architecture report

D2.4 Final requirements and specifications for pilot demonstrators

Work Package 3 – Photorealistic 3D model capture and animation
Lead: Fraunhofer

Work package 3 addresses the capture of the 3D human models for photo-realistic rendering as well as the capturing of real body and face motion for animation. In order to create realistic appearance als for detailed regions – e.g. finger animations-, video-based approaches will be used that synthesize the output from real video footage together with simple 3D body models. The work package covers the creation of hybrid human body models from sample captures, tools for the creation of animation samples as well as methods for an efficient rendering and animation.

Deliverables available:

D3.1 Initial Report on 3D model capture and rendering

Work Package 4 – Signal processing for sign-language interpretation
Lead: Surrey

The challenging task of translation from written language to that of Sign is addressed in work package 4. As this is currently an unsolved problem, CONTENT4All will limit demonstration to a controlled domain of discourse. The objective of this work package are:

to build a deterministic grammar model which can be used to map spoken language to the multi-channel constructs of Sign,

to analyse the data captured in phase 1 of CONTENT4All to identify the components of Sign that correspond to submit components of Sign,

to generate new sequences of Sign that are plausible translation of spoken text within the selected domain of discourse.

Work Package 5 – System integration, validation, large pilot set-up and user testing
Lead: Fincons 

This work package aims at integrating the outcomes from work package 2-4 to provide a comprehensive system validating the CONTENT4All concept that could also be adopted by a broadcaster. The main objective of this work package is the implementation, integration, validation and execution of the 2 pilot technology demonstrators developed in CONTENT4All. As such, in addition to developing the roadmaps for realizing the pilots, another objective of this work package to identify weaknesses in the developments an to provide feedback to work package 2, 3 and 4.

Deliverables available:

D5.1 Pilot 1 demonstrator architecture, integration plan, and evaluation methodologies

D5.3 Pilot 2 demonstrator architecture, integration plan, and evaluation methodologies

Work package 6  –   Dissemination, communication, standardization, exploitation, business planning
Lead: SwissTXT

The primary objective of this work package is to communicate the innovations in CONTENT4All, and to disseminate the outcomes such at the software developed and datasets generated during its course. The second objective of the work package will be to explore possible business models for the technologies, in part and as a whole, and to develop mechanism to exploit these technologies for innovative applications in the future.