By the end of the project the consortium made important progress towards its objectives in the following ways:
Technical partners have worked together to implement the MediaVerse decentralised framework of digital asset management systems. The final release of the MediaVerse platform was achieved, presenting a report containing the updated system architecture based on the analysis of user requirements, as well as the final results from components integration and testing. Seven MV nodes have been installed by different partners.
Thorough legal and technical analysis of content IPR issues has been carried out, presenting a functional and technical description of the content discovery system, copyright compatibilities and blockchain infrastructure inside MediaVerse. Moreover, the legal and ethical compliance of the project offers legal support for IPR management.
The latest advances in deep learning for media annotation have been tested for text, images, videos, and 3D content. Several analysis models have been deployed as part of a Media Annotation service. Software components, as well as improvements and additions in the Media Annotation Service and the adopted retrieval and recommendation technologies were described. Furthermore, a Media Asset Annotation and Management (MAAM) platform, evolved as a fork out of the main MediaVerse platform.
Accessibility-by-design in content creation and enrichment has been supported by the accessibility toolset (i.e. automatic subtitle generation, editing and semantic enrichment), outlining the progress made in creating an accessible platform, ensuring equal access to information and content for all. Many of these features have been tested in pilots and the evaluation results have been reported.
The Fader and VRodos XR authoring tools have been integrated to the MediaVerse platform to the extent that was deemed necessary by the pilots and user requirements.
In the domain of social analytics, significant changes were announced by Twitter/X in 2023, including limitations to API access and pricing. To counter this, the Social Analytics Engine was updated with new functionalities, including the creation of PDF reports, improvements in trending and alerting, the introduction of manual sentiment editing, and the creation of a new web application for mass correction of entity extractions.
Large-scale pilot activities have successfully taken place for all Use Cases and the final evaluation results were reported.
Consortium partners performed a variety of dissemination and communication activities, including presentations to events, trainings, clustering activities, blog posts, and publications. Also, an elaborate version of the exploitation plan, along with market analysis activities, have been delivered.