XReco

XR mEdia eCOsystem

While media organisations increasingly support non-linear experiences for the consumer, those are still limited to single channels and media domains. Although several media organisations have recently succeeded in breaking data silos, data sharing is mostly limited to the organisation. Thus, there are challenges for producing content feeding multiple channels with different granularities and structures, mainly related to data discovery, management and (re-)use. XRECO will create a new data-driven ecosystem for the media industry, focusing on facilitating data sharing, search and discovery and supporting creation of news and entertainment content, in particular, the creation and (re-)use of location-related 2D and 3D assets and the creation of XR experiences. The ecosystem core, represented by a Neural Media Repository (NMR), will foster inter-organisation content sharing and provide increased access to content for media creators, considering novel data monetization and rights management policies. A set of AI-based media transformation services are built around the NMR to produce novel media- and XR experiences, including 3D neural reconstructions, neural based device localisation, image stitching, de-/re-lighting and holoportation. The developed technology will be validated in use case scenarios for (i) the news media for XR-based broadcasting and automatic and customized multitarget news publishing, and for (ii) location-based information and entertainment content, with applications in tourism and the automotive industry.

Project domains

Sectors
Automotive
Entertainment
Tourism
Interaction & Enabling Technologies
Extended Reality

Publications

Title Type of output
Efficient Few-Shot Incremental Training for Landmark Recognition Conference proceedings
Image Valuation in NeRF-Based 3D Reconstruction Book chapters
Complete Convolutional Neural Networks Environment for Computer Vision Problems With Nvidia Drive AGX Xavier Book chapters
Synthetic Football Sprite Animations Learned Across the Pitch Book chapters
Descriptor Impact on Multimodal 3D Retrieval Conference proceedings
A Dataset and Metric for Textual Video Content Description Conference proceedings
Urban Scene Removal and Completion Conference proceedings
Analysis of Objective 3D Mesh Quality Metrics for Cultural Heritage Conference proceedings
Untethered Real-Time Immersive Free Viewpoint Video Conference proceedings
Multimodal Understanding: Investigating the Capabilities of Large Multimodal Models for Object Detection in XR Applications Conference proceedings
Volumetric Video Reconstruction and Communications: Toward a New Era of Interactive and Immersive Social Virtual Reality (VR) Experiences Conference proceedings
Semi -Automated Digital Human Production for Enhanced Media Broadcasting Conference proceedings
Multimedia Information Retrieval in XR Conference proceedings
3DMSE: An Interactive 3D Media Search Engine Conference proceedings
Is Real-time Deep Learning-based Monocular Depth Estimation accurate for Multi-Camera Setups? Conference proceedings
Analysis and Development of Deep Learning Depth Estimation Techniques for Volumetric Capture and Free Viewpoint Video Conference proceedings
Real-Time Free Viewpoint Video for Immersive Videoconferencing Conference proceedings
Multimodality in Media Retrieval Conference proceedings
Exploring Image Search on Quantum Computing Systems Conference proceedings
Towards a Universal Query Representation for Multimodal Information Retreival Conference proceedings
Multimedia Retrieval in and for XR Conference proceedings
Enabling Domain Experts to Train Efficient Few-Shot Incremental Landmark Recognition Conference proceedings
XReco Platform and RAI News Media Demonstrator Conference proceedings
Open-Source Multimedia Retrieval with vitrivr-engine Conference proceedings
Ubervvald: Advanced Object Detection Library for Optimizing Complex Convolutional Neural Networks (CNNs) Conference proceedings
Nostrils and Mouth Detection for Drivers Using Convolutional Neural Networks with Automatically Generated Ground Truth Data Conference proceedings
Ground Truth Data Generator in Automotive Infrared Sensor Vision Problems Using a Minimum Set of Operations Conference proceedings
A new Retrieval Engine for vitrivr Conference proceedings
Improving Query and Assessment Quality in Text-Based Video Retrieval Evaluation Conference proceedings
Real-Time Layered View Synthesis for Free-Viewpoint Video from Unreliable Depth Information Conference proceedings
Free-form Multi-Modal Multimedia Retrieval (4MR) Conference proceedings
Multimedia Retrieval in Mixed Reality: Leveraging Live Queries for Immersive Experiences Conference proceedings
Mining Landmark Images for Scene Reconstruction from Weakly Annotated Video Collections Conference proceedings
Synthetic Football Sprite Animations Learned Across the Pitch. Conference proceedings
Subjective Evaluation of Dynamic Point Clouds: Impact of Compression and Exploration Behavior Conference proceedings
Multimodal 3D Object Retrieval Conference proceedings
Interactive Multimodal Video Search: An Extended Post-Evaluation for the VBS 2022 Competition Peer reviewed articles
Multimedia Systems Peer reviewed articles
An Assessment of the Stereo and Near-Infrared Camera Calibration Technique Using a Novel Real-Time Approach in the Context of Resource Efficiency Peer reviewed articles
Data as Remuneration in Digital Copyright Licensing: Some Reflections on the Concept of ‘Appropriate and Proportionate Remuneration’ Under Art. 18 EU Directive 2019/790 in the Data Era Conference proceedings
Secure, Dynamic and Uncomplicated Licensing of Movies on a Blockchain Infrastructure Conference proceedings
IEEE Access Peer reviewed articles
Delay Threshold for Social Interaction in Volumetric eXtended Reality Communication Peer reviewed articles
VERGE in CBMI2024 Conference proceedings
VERGE in VBS 2024 Conference proceedings

Do you want to learn more or discuss about this project?
Visit the Openverse Wiki page for XReco

Disclaimer and intellectual property acknowledgement: the materials showcased on this page have been produced by third parties, and the OPENVERSE consortium accepts no liability for any inaccuracies therein. Responsibility for the content of each asset rests solely with its original creator. The inclusion of these outputs on this page does not, in any way, constitute an endorsement on the part of the OPENVERSE consortium. All intellectual property rights relating to the content published here remain vested in their respective owners. This page should be understood simply as a practical point of access.