CORTEX2 innovators: CDLPG's 1st progress update
Gesture control is one of the most intuitive ways to interact in extended reality (XR) — but making it work reliably across different users, devices, and platforms is a challenge. That is where CDLPG comes in. Supported by CORTEX2, this project is developing a scalable system for personalised gesture recognition that integrates directly with the CORTEX2 framework.
We spoke with the team to learn more about their vision, progress, and next steps.
Q: What is CDLPG, in one sentence?
A: The CDLPG project (Co-development of a Dynamic Library of Personalised Gestures), led by Sensorama Lab, aims to create a dynamic, scalable gesture recognition system integrated with CORTEX2, enabling personalised hand and body gestures to trigger predefined actions.
Q: What problem are you solving, and what makes your solution unique?
A: We are tackling a key challenge in the AR/VR space: creating a gesture recognition module that can adapt to various use cases within the CORTEX2 framework. Most existing systems struggle with recognising a wide range of specific gestures — especially when performed differently by different users. That impacts consistency and usability.
We are also working to overcome the technical challenges of integrating gesture recognition with diverse AR/VR hardware. Our solution is designed to work across platforms and devices, making it more accessible and reliable.
Q: What are CDLPG’s main objectives?
- Develop gesture recognition: Accurately capture and interpret hand and body gestures.
- Integrate with CORTEX2: Ensure compatibility with AR/VR hardware and the CORTEX2 framework.
- Enable gesture-to-action mapping: Link gestures to predefined actions within applications.
CORTEX2 support programme progress
Q: What have you achieved so far?
A: We have developed a system that connects input data with gesture recognition components and created a functional prototype where gestures are processed and recognised. Here is a short demo of the CDLPG prototype:
https://youtube.com/shorts/uYLxww2QFA0
An operational gesture registration system has been built, capable of recognising at least seven distinct gestures. The system has been integrated with one of the CORTEX2 pilots, allowing compatibility with its framework. Additionally, we have enabled gesture-to-action mapping and integrated a dynamic action mapping component within the CORTEX2 pilot.
We expect that our solution will bring about the following impact:
- Enhanced user interaction: By implementing this solution, we aim to transform user engagement within AR/VR realms. Our approach centers on integrating seamlessly intuitive, gesture-driven interfaces, redefining the way users navigate and interact within these digital spaces. This shift towards more organic and user-friendly controls is set to elevate the overall immersion and accessibility of AR/VR experiences.
- Increased versatility of AR/VR applications: With a dynamic library of gestures, AR/VR applications can be tailored to a wide range of uses, from collaboration to professional training and education.
Q: How is participating in CORTEX2 supporting CDLPG?
A: Participating in the CORTEX2 program has been very valuable for our team. As a Ukrainian company, being part of a joint European XR platform allows us to collaborate with international partners and contribute to a shared project. This opportunity enables us to exchange knowledge and experience with our European counterparts, helping us improve and grow. Working within CORTEX2 also supports the development and integration of our gesture recognition system, with valuable input and feedback from the wider network. We are proud to be part of this collaborative effort.
Q: What is next for CDLPG?
A: Next, we will focus on testing, refining, and promoting the system. We will develop a checklist to test the system's reliability and conduct thorough evaluations. Focus groups and usability studies will be organised to gather user feedback, which will help us improve performance and usability. Detailed documentation of the software module will be prepared, including mentor feedback, and published on GitHub. Finally, we will create a promotional video to showcase the project and share its progress.
Learn more about CDLPG and stay updated on its progress!
Want to explore more XR innovation? Browse all our supported projects on the CORTEX2 website:
Open Call 1 winners - Open Call 2 winners
CORTEX2 innovators: ARY's 1st progress update
We caught up with the team behind ARY, one of our Open Call 1 winners. They are developing a solution to a key challenge in extended reality (XR): enabling multiple XR devices to share the same spatial understanding of their environment — without the need for markers or scanning.
Let’s hear from the team about what they are building and how their journey is unfolding.
Q: What is ARY, in one sentence?
A: ARY, the AR Media, is an innovative solution to bring spatial awareness to XR devices — led by ARY.
Q: What problem are you solving, and what makes your solution unique?
A: XR devices typically operate within their own isolated environments. To ensure an object appears at the same physical location across multiple devices, it must possess spatial awareness.
ARY introduces a wireless solution that requires no markers or environment scanning. This approach leverages radio technology to enable accurate positioning in indoor environments.
Q: What are ARY’s main objectives?
A: We aim to design a beacon-based system that brings spatial awareness to connected XR devices. It is cross-platform and works on any Bluetooth device with the help of a small add-on.
CORTEX2 support programme progress
Q: What have you achieved so far?
A: We have successfully designed hardware beacons that provide spatial awareness to other devices. These are the foundation of our solution.
Q: How is participating in CORTEX2 supporting ARY?
A: CORTEX2 has given us a unique opportunity to develop and fund this technology. It offers a valuable platform for working on real-world use cases. We also benefit from project management support and access to mentors, which helps us grow the technology further.
Q: What are your next steps within the programme?
A: We will be developing Unity modules to support integration within CORTEX2. We will also build an add-on that makes any device not natively supported by ARY compatible with our system.
Learn more about ARY and stay updated on its progress!
Want to explore more XR innovation? Browse all our supported projects on the CORTEX2 website:
Open Call 1 winners - Open Call 2 winners
BeyondXR: Shaping the Future with XR, AI and Robotics Across Education, Industry and Emerging Technologies.
A new project cluster, BeyondXR, has been established, bringing together several pioneering EU projects focused on XR (Extended Reality) and emerging technologies. This initiative aims to reshape learning and training by integrating advanced technologies such as XR, AI, and robotics into transformative educational solutions, address broader applications of XR and emerging technologies, including cross-industry collaboration, immersive media, manufacturing innovation, creative and cultural advancements, and the social acceptance of XR systems.
By fostering synergies among projects, BeyondXR seeks to revolutionise how XR and AI technologies are applied across various sectors to drive innovation and societal impact.
The cluster comprises some of the most innovative projects in the field, including:
- CORTEX2: revolutionising cross-industry collaboration through a cutting-edge XR platform, designed to break down barriers and facilitate seamless interaction across industries.
- XR5.0 Project: The XR5.0 project aims to build, demonstrate, and validate a novel Person-Centric and AI-based XR paradigm tailored to the requirements and nature of I5.0 applications.
- XR4ED: building a European reference platform for EdTech and XR communities, enhancing learning and training with XR technologies, and creating valuable educational opportunities.
- MASTER: advancing robotics training within the manufacturing sector, providing learners with hands-on experience and skills needed to thrive in an automated world.
- HECOF: revolutionising teaching in higher education through AI-driven, personalised, adaptive learning. Focused on Chemical Engineering, it leverages digital data to enhance student assessment and learning methods.
- augMENTOR: revolutionising education by integrating emerging technologies into a pedagogical framework focused on developing both basic and 21st-century skills. Using an AI-boosted toolkit, the project will leverage machine learning algorithms to provide personalised learning pathways tailored to individual learners’ characteristics and preferences.
- XR2Learn: aims to establish the cross-border creation of human-centric XR applications in education. The project will deliver its one-stop-shop platform, organised as a Digital Innovation Hub, for all actors involved in the XR-based educational applications supply chain, aimed at enhancing training in manufacturing and distance learning scenarios.
- SERMAS: lays the foundations of next-generation XR systems, looking at how people relate and interact with the technology. Transform your XR experience with SERMAS. Embrace social acceptance and discover our concepts, techniques, and tools that will help you discover the full potential of XR systems through coherent design, implementation, and deployment.
- TRANSMIXR: aims to ignite the immersive media sector by enabling new narrative visions through the development and adoption of XR and Artificial Intelligence (AI) technologies. The project focuses on advancing social XR and AI technologies for application within media production, delivery, and consumption, targeting the creative and cultural sectors.
For more than six months, the BeyondXR cluster has organised monthly meetings to plan future joint events, promote XR activities, and gather feedback and recommendations from the participating projects. This collaborative approach ensures continuous improvement and alignment of efforts to maximise impact. By leveraging emerging technologies, these projects aim to drive meaningful changes, fostering innovation and building the future of learning in the XR environment.
Follow our updates on social media through the #BeyondXR hashtag!
Announcing the winners of the CORTEX2 Open Call #2
We are excited to introduce the CORTEX2 Open Call #2 winners! An impressive group of innovators and visionaries committed to shaping the future of extended reality (XR) technologies and developing inclusive, efficient, and immersive experiences. After a competitive selection process, 10 groundbreaking projects were chosen, representing 12 beneficiaries from 8 countries, including 7 SMEs, 2 startups, and 3 research organisations (ROs).
The selected projects will contribute to the co-development of the CORTEX2 platform, tackling key challenges in XR, including sign language translation, immersive 3D collaboration, and anonymisation of data in XR environments.
They will demonstrate XR’s transformative potential in improving remote collaboration, communication, and training across industries.
Meet these projects and explore how they are advancing the XR landscape:
INTERACT: Inclusive Networking for Translation and Embodied Real-Time Augmented Communication Tool with Sign Language Integration
- Lead organisation: DASKALOS-APPS (France)
- Topic: Embodied Avatar
Objective: Develop AI-powered avatars for real-time sign language translation in augmented reality (AR) to support multilingual and hearing-impaired participants in business meetings.
INTERACT will leverage the CORTEX2 architecture to deliver scalable, sentiment-aware interactions in teleconferencing settings. It will improve accessibility for deaf and hard-of-hearing individuals, promoting inclusion in business meetings.
VISIXR: Vision AI for XR
- Lead organisation: ZAUBAR (Germany)
- Topic: Smart Generator
Objective: Develop a cutting-edge Smart Generator tool enabling real-time, AI-driven modification and understanding of 2D/3D assets within real-time 3D environments, integrating these into the CORTEX2 ecosystem.
VIRTEX: Virtual Experience Creation Platform for Seamless Integration of CORTEX2 Services in Industrial and Commercial XR Applications
- Lead organisations: MetaMedicsVR (Spain), Ludwig-Maximilians-Universität (Germany)
- Topic: Virtual Experiences Editor
Objective: Democratising XR technologies and enhancing remote collaboration across diverse industrial and commercial sectors.
To achieve this mission, VIRTEX aims to develop an innovative editor that integrates CORTEX2 services into VR/Web 3D applications. The project responds to the need for more effective remote collaboration and training tools, underscored by recent pandemics and wars.
NITRUS: Nonverbal Interactive Tool for gesture Recognition in mUltiperson Scenarios
- Lead organisation: Logicmelt Technologies SL (Spain)
- Topic: MPRR (Multi-Person Reaction Recognition)
Objective: Use state-of-the-art AI models to create a gesture recognition solution that enriches the interaction of people’s digital experiences.
Its key goals are to gather and create unbiased gesture datasets, benchmark and train AI algorithms for gesture recognition, optimise and develop pipelines for the real-time execution of the algorithms, and integrate them into the CORTEX2 framework.
NODAV: Next-Generation 3DGS Optimisation And Visualisation
- Lead organisation: i2CAT Foundation (Spain)
- Topic: Gaussian-splatting-based reality capture for VR
Objective: Advance Gaussian Splatting (GS) technology to enable high-quality reconstruction of static scenarios.
The project aims to tackle challenges in data efficiency by implementing pruning, data reduction, and compression solutions for improved transmission and storage of GS reconstructions. Additionally, NODAV seeks to unlock the potential of GS for real-time rendering in Unity across VR, mobile, and web applications. By addressing these critical aspects, NODAV enhances GS technology’s accessibility and scalability, paving the way for seamless integration into modern interactive and immersive platforms.
SAME-XR: Scalable Asset Management and Conversion Engine for XR Development
Lead organisation: Nara Eğitim Teknolojileri AŞ (Turkey)
Topic: 3D Model Database
Objective: Develop a comprehensive asset management tool to streamline 3D asset development workflow and integrate it into the CORTEX2 platform.
SAME-XR will simplify asset management for XR developers, saving time and improving workflows.
Some of its key innovations will be:
- Cloud-native 3D asset database with a comprehensive toolset
- Centralised library for 3D assets
- Unity GUI editor tool and add-ons for 3D software
- Asset discovery through integration with external databases
- Secure asset management with version control
- Support for multiple file formats and optimisation tools
- Web-based interface with 3D and AR preview capabilities
- Multi-User XR collaboration/prototyping.
PETER: Preserve Emotions in Translations for Extended Reality
- Lead organisations: Università degli Studi di Cagliari, R2M Solution S.r.l. (Italy)
- Topic: Real-time voice translation
Objective: Develop a quasi-real-time voice-to-voice translation system that preserves the emotional tone of XR applications.
The project will deliver fast, emotionally aware translations to enhance engagement and communication effectiveness and reduce misunderstandings while using XR technologies.
XR-CARE: Extended Reality Collaborative Anonymisation for Remote Healthcare
- Lead organisation: Logimade Lda (Portugal)
- Topic: Anonymising meetings content for privacy-free data storage
Objective: Anonymise XR teleconference data while preserving usability.
The project will anonymise multiple data types, including high-definition video, audio, and physiological data, without compromising personal information and preserve the rest of the content intact and usable.
VEM: Virtual Encrypted Meetings
- Lead organisation: Simplito sp. z o. o. (Poland)
- Topic: Open – Own project idea
Objective: Enhance data privacy and security in VR communication platforms.
The project's key features are comprehensive end-to-end message encryption (server & user level) and full data protection, ensuring secure VR communication for users.
R3in3D: RGBD Real-Time Representations of Humans and Objects in 3D
- Lead organisation: TNO, Netherlands Organisation for Applied Scientific Research (Netherlands)
- Topic: Open – Own project idea
Objective: Provide immersive 3D representation of humans and objects for XR applications.
R3in3D aims to redefine collaborative business meetings and remote training experiences in XR, fostering natural and engaging communication between participants and improving decision-making, knowledge retention, and skill acquisition.
The project will develop tools for realistic 3D capture and rendering modules for Unity and WebXR. Thus, it will allow the building of holographic communication pipelines while making the most out of next-generation networks (5G and 6G) and “network slices." These customisable pipelines allow the 3D representation of a real-world working station for a novel, immersive way of training in XR environments and business meetings.
Next step: The CORTEX2 Support Programme
Over the next 9 months, these projects will receive funding, resources, and expert guidance to advance their groundbreaking solutions and bring tangible results to life—all while they help shape the CORTEX2 platform, ensuring it evolves into a robust, scalable solution.
We look forward to showcasing their progress and achievements as they work to shape the future of XR.
Stay tuned for updates on their progress!
Follow us on X and LinkedIn
Subscribe to our newsletter
CORTEX2 Publication: G3FA: Geometry-guided GAN for Face Animation
Our paper, “G3FA: Geometry-guided GAN for Face Animation”, was presented at the British Machine Vision Conference 2024.
Abstract
Animating human face images aims to synthesize a desired source identity in a natural-looking way mimicking a driving video’s facial movements. In this context, Generative Adversarial Networks have demonstrated remarkable potential in real-time face reenactment using a single source image, yet are constrained by limited geometry consistency compared to graphic-based approaches. In this paper, we introduce Geometry-guided GAN for Face Animation (G3FA) to tackle this limitation. Our novel approach empowers the face animation model to incorporate 3D information using only 2D images, improving the image generation capabilities of the talking head synthesis model. We integrate inverse rendering techniques to extract 3D facial geometry properties, improving the feedback loop to the generator through a weighted average ensemble of discriminators. In our face reenactment model, we leverage 2D motion warping to capture motion dynamics along with orthogonal ray sampling and volume rendering techniques to produce the ultimate visual output. To evaluate the performance of our G3FA, we conducted comprehensive experiments using various evaluation protocols on VoxCeleb2 and TalkingHead benchmarks to demonstrate the effectiveness of our proposed framework compared to the state-of-the-art real-time face animation methods. Our code is available at github.com/dfki-av/G3FA.
Authors
Alireza Javanmardi, Alain Pagani, Didier Stricker
Read the full publication
Access all our CORTEX2 publications.
CORTEX2 Publication: CoVA: A Virtual Assistant for Virtual Reality Conferencing
Our paper, “CoVA: A Virtual Assistant for Virtual Reality Conferencing”, was presented at the Proceedings of the 21st EuroXR International Conference (EuroXR 2024).
Abstract
Teleoperation and collaboration are among the key pillars of business work, where services and demand are spread over a very wide economic market, such as the European Union. What characterizes their importance is not only remote exchanges, but also the ability to intervene or assist people without having to travel, as in the example of machine maintenance and repair. This is made possible by the integration of IoT and artificial intelligence into this vast technological field, which is further accentuated by extended reality.
CORTEX² project is in line with this vision to bridge the divide between widespread videoconferencing tools and state-of-the art XR-based solutions, democratizing the uptake of next-generation Extended Reality tele-cooperation among many industrial segments and SMEs.
One of the many promising developments in this space is the integration of AI-based conversational agents within XR environments (Reiners et al., 2021). When combined with XR applications, virtual agents can facilitate real-time collaboration, information retrieval, and task automation. However, this combination presents several challenges. In multi-party dialog contexts, where participants interact simultaneously, AI conversational agents must accurately handle real-time speech recognition and response generation (Clark et al., 2019). Achieving this at scale requires the use of highly efficient tools and models to minimize latency and ensure smooth, uninterrupted conversations, particularly when managing multiple users concurrently.
In the rest of this paper, we will illustrate the contribution of artificial intelligence to extended reality through an XR video conferencing application, where an AI virtual assistant plays a critical role in business meetings.
Authors
Alexis Lombard, Yazid Benazzouz, Galo Castillo López, Gaël de Chalendar, and Jean Pierre Lorré
Read the full publication
Access all our CORTEX2 publications.
CORTEX2 Publication: Augmented telecommunication in factory setting
Our paper, “Augmented telecommunication in factory setting”, was presented at the Proceedings of the 21st EuroXR International Conference (EuroXR 2024).
Introduction
As a result of the COVID-19 pandemic, the possibility of remote work or “Home Office” has been normalized by many companies. Although the existing telecommunication tools are sufficient for many mainstream tasks, they lack the capabilities of 3D interaction which is necessary for complex tasks which require physical presence for engaging in problem solving.
As an early implementation of collaborative 3D communication tools, Microsoft (Chen et al., 2025) developed the HoloLens system, which introduces a novel interaction model for supporting collaboration between a head-mounted display (HMD) user and remote participants. The HoloLens allows remote companions to join the AR space by hitching onto the view of the primary HMD user through Skype-enabled devices, such as tablets or PCs. This system facilitates asynchronous interaction in a shared 3D space with digital objects, allowing remote parties to contribute to tasks and have their inputs reflected back to the primary user in real-time, thus enabling new scenarios for remote collaboration.
Further advancements in remote collaboration systems have focused on complex tasks like environmental pollution analysis, which require expertise from multiple fields. One such system was designed by Mahmood et al. (2019). It uses mixed reality to support co-presence and collaborative analysis, demonstrating improved remote analysis through shared user and data spaces. Drey et al. (2022) explored how the benefits of pair-learning and virtual reality (VR) can be combined by comparing symmetric systems, where both peers use VR, and asymmetric systems, where only one peer uses VR and the other uses a tablet. They found that the symmetric system significantly enhanced presence, immersion, and reduced cognitive load, which are important for learning. However, both systems resulted in similar learning outcomes, demonstrating that both symmetric and asymmetric setups are effective for co-located VR pair-learning.
In industrial and technical settings, operating machinery often requires assistance or training that can be difficult to acquire with traditional documentation or voice/video calls alone. These methods often fail to convey spatial relationships, leading to miscommunication and repeated explanations. To address these challenges, we start by 3D scanning the machines and environments ahead of time, to have them available when running the application. The technician on site is assisted by a remote expert, with the option for additional observers, using multi-device support. Depending on available hardware, participants join the session through their respective devices (PC, XR headset), with the software adapting to features like webcams and tracking.
In this setup, the expert views a virtual representation of the object or environment and can track the technician’s pose to better understand what they are looking at. The expert can place and manipulate 3D annotations in the scene and provide additional guidance via voice and video. The technician sees the actual scene through a webcam or XR headset with the expert’s annotations superimposed, matching the 3D position. This setup enables efficient collaboration between the expert and technician to solve complex problems more effectively.
While existing systems such as HoloLens and mixed reality platforms focus on immersive experiences, our approach is tailored to the industrial environment. It addresses the current issue of heterogeneous hardware availability and usage, allowing flexibility through multi-device integration and adapting to the hardware on hand.
Authors
Narek Minaskan, Bastian Krayer, Alain Pagani, and Didier Stricker
Read the full publication
Access all our CORTEX2 publications.
CORTEX2 Publication: AR Assistant for Pruning of Grapevines and Fruit Trees
Our paper, “AR Assistant for Pruning of Grapevines and Fruit Trees”, was presented at the Proceedings of the 21st EuroXR International Conference (EuroXR 2024).
Introduction
Winter pruning in orchards is an essential but labor-intensive and time-consuming task, traditionally done manually to shape future growth by removing unwanted branches. Workers use various techniques to optimize yield, fruit quality or disease resistance. However, labor shortages and the need for professional training pose challenges for farmers. To address this, we offer support to make pruning more accessible to a broader segment of the workforce.
At the same time, computer vision in outdoor environments is complex due to varying lighting, weather, unique plant shapes, occlusions and similarity between foreground and background plants. Multiple studies have addressed the problem of pruning grapevines and other fruit trees (Amatya, 2016; Botterill, 2017; Gentilhomme, 2023; Tong2023). These studies either focus on the entire automation pipeline (Botterill, 2017; Fourie, 2021) or separate steps, for instance, branch detection (Amatya, 2016; Zhang, 2018), reconstruction and skeletonization (You, 2022; Feng, 2024), and cut position localization (Marset, 2021). The recent approaches (Fourie, 2021; Gentilhomme, 2023) prove the general feasibility of automated pruning systems but do not address real-world challenges like outdoor conditions and complex, occluded plant structures. Effective pruning systems for fruit trees require accurate spatial information. Several studies have highlighted challenges in capturing thin structures using laser scanners or 3D cameras, often requiring additional refinement or proper initial registration, which can be time-consuming (Tagarakis 2013; Medeiros, 2017). To achieve a balance between cost and benefit, affordable methods for 3D reconstruction need to be explored. In contrast to similar studies that use 3D sensors for apple trees (Majeed, 2018; Tong, 2023), we explore the potential of image-based approaches suitable for an augmented reality (AR) pruning assistant on a mobile device.
In this work, we address the mentioned challenges and present an AR assistant that enables inexperienced workers to carry out pruning for grapevines and reduces the size of the cut wounds, making the plants more resilient to fungal infections and promoting rich and healthy yield. We further apply this concept to other fruit trees, such as apple and peach trees, and highlight the improvements made in this direction. Our contributions can be summarized as follows. First, we present a pipeline that extracts 3D and semantic information from a video of a plant and outputs pruning suggestions using both traditional and deep-learning methods (Vid2Cuts). Second, we introduce a mobile AR application to display the results to the user. Third, we extend the pipeline for other fruit trees that are more challenging compared to grapevines due to their more complex 3D structure and larger size.
Authors
Mariia Podguzova, Simon Häring, Jamiu Ojeleye, Stephan Krauß, and Didier Stricker
Read the full publication
Access all our CORTEX2 publications.
CORTEX2 will participate at EuroXR Conference
We are thrilled to announce that CORTEX2 will participate in the 21st EuroXR International Conference (EuroXR 2024), a leading event in the fields of virtual reality (VR), augmented reality (AR), and mixed reality (MR). This year, the conference is co-organized by the Institute of Communication and Computer Systems (ICCS) of the National Technical University of Athens (NTUA) and will take place in Athens, Greece, from November 27 to 29, 2024. It will provide an excellent opportunity for us to showcase our advancements in extended reality (XR) and engage with the broader XR community.
Every year, EuroXR attracts a diverse audience, including researchers, developers, industry leaders, and policymakers, eager to learn about the latest developments in VR, AR, and MR, and passionate about the transformative potential of XR technologies.
This year’s conference will highlight several EU-funded projects, including CORTEX2, SUN-XR, DIDYMOS-XR, THEIA-XR, SHARESPACE and LUMINOUS, along with initiatives from the Alliance4XR project.
Join us in shaping the future of XR at EuroXR 2024
We will seize this opportunity to connect with the XR community, share our work, and further strengthen our collaborations. We look forward to seeing many of you in Athens!
Learn more about EuroXR 2024 and secure your spot.
CORTEX2 Publication: Leveraging Discourse Structure for the Creation of Meeting Extracts
We have published the paper “Leveraging Discourse Structure for Extractive Meeting Summarization”.
Abstract
We introduce an extractive summarization system for meetings that leverages discourse structure to better identify salient information from complex multi-party discussions. Using discourse graphs to represent semantic relations between the contents of utterances in a meeting, we train a GNN-based node classification model to select the most important utterances, which are then combined to create an extractive summary. Experimental results on AMI and ICSI demonstrate that our approach surpasses existing text-based and graph-based extractive summarization systems, as measured by both classification and summarization metrics. Additionally, we conduct ablation studies on discourse structure and relation type to provide insights for future NLP applications leveraging discourse analysis theory.
Authors
Virgile Rennard, Guokan Shang, Michalis Vazirgiannis, Julie Hunter
Read the full publication
Access all our CORTEX2 publications.
Announcing the results of CORTEX2's second Open Call
We are excited to announce that we have achieved great success with the results of our second Open Call, which launched on June 13 and closed on August 15, 2024. We received 91 applications from 25 countries, covering all the call topics. Thanks to all applicants for your interest in joining our journey to democratise extended reality solutions for easy, efficient, and accessible remote collaboration, and good luck!
This opportunity is aimed at extended reality (XR) innovators (from tech startups/SMEs to researchers) to co-develop our innovative XR teleconference platform. They will introduce new modules and features, enhancing the platform's functionalities and opportunities.
Selected beneficiaries will receive funding (up to 100,000 EUR per project) and access to our 9-month support programme, which includes tailored guidance and support, access to tech and business experts, capacity building on CORTEX2 and XR technologies & trends, and resources to facilitate the integration and understanding of the CORTEX2 platform.
The CORTEX2 Open Call #2 results
Our second call has attracted a large number of applications from diverse origins, with 91 submitted out of 149 started — 19 proposals, including a consortium of 2 organisations.
Regarding interest distribution by topic, the open topic has received the most applications at 24, followed by Virtual Experiences Editor with 16 applications and Embodied Avatar with 11 applications.
Distribution of applications submitted to the Open Call #2 topics
- Open topic: 24 applications
- Virtual Experiences Editor: 16 applications
- Embodied Avatar: 11 applications
- MPRR (Multi Person reaction Recognition): 9 applications
- Real-time voice translation: 9 applications
- Gaussian-splatting-based reality capture for VR: 8 applications
- 3D model database: 7 applications
- Аnonymizing meeting's content for privacy-free data storage: 5 applications
- Smart generator: 2 applications
The evaluation process is set to conclude by September 2024, and results are expected by the end of the month.
Meet our XR innovators from Open Call #1
We recently welcomed our Open Call #1 winners to CORTEX2! Twenty teams of exceptional professionals with innovative and diverse solutions that align with our mission to accelerate and democratise XR technology across Europe.
These teams will co-develop our cutting-edge XR platform to create value-added services and engage new use cases to demonstrate its adaptability in different domains. They will receive funding and mentorship to bring their visionary concepts to life.
We are incredibly excited about their potential and the impact they will make in the XR field, and we look forward to seeing how they contribute to the CORTEX2 ecosystem!
Stay tuned for more updates on the progress of these groundbreaking projects and the results of our Open Call #2.
Follow us on X and LinkedIn
Subscribe to our newsletter
[RECORDING] CORTEX2 Open Call 2 Webinar 2: Application topics
On 31 July 2024, we held our second info webinar about our Open Call 2. In it, our technical colleagues presented the topics to apply to, covering challenges, requirements, expected outcomes, and what (technical) support and resources we will provide to the winners during our 9-month support programme.
Now that we have developed the backbone and specific features of our innovative extended reality (XR) teleconference platform, we are looking for partners — companies (tech startups/SMEs) and research institutions (universities, NGOs, foundations, associations) — to collaborate with us on further developing it, providing new modules and features to expand its functionalities.
Applicants will become eligible to receive up to €100,000 and access our 9-month support programme. This includes tailored guidance and support, as well as access to technology and business experts, capacity building, and resources to facilitate the integration and understanding of our platform.
The open call topics
As an applicant, you should choose one of these topics to apply to. If you don’t find a suitable one, you can also apply for an open topic aligned with the CORTEX2 framework and objectives.
- Embodied Avatar
- Smart generator
- Virtual Experiences Editor
- MPRR (Multi Person reaction Recognition)
- Gaussian-splatting-based reality capture for VR
- 3D model database
- Real-time voice translation
- Аnonymizing meeting’s content for privacy-free data storage
- OPEN TOPIC: Submit your own project idea
The open call will be open until 15 August 2024 at 17:00 CET.
Check the Open Call 2 website, carefully review the call documents and recording below, and prepare to make a successful application.
Apply now!
https://www.youtube.com/watch?v=JlkiyNHxRMo