NGA – Next Generation of Avatars

More details can be found here: project.inria.fr/nga2025

Recent advances in deep learning, neural representations, and physics-based character simulation… have contributed to a breakthrough in avatar design, simulation, control and rendering. Computer vision now enables detailed and accurate pose and shape capture with affordable systems (e.g. a single RGB camera) in real-time, and to build controllable representations of specific users instead of anonymous virtual humans. In particular, advances in image-based rendering, such as NERF or Gaussian Splatting, also offer new challenges to control high quality avatars, opening new questions about avatar control, and its impact on human experience in VR. Novel view synthesis based on sparse views of the user may be a promising alternative to traditional character simulation, especially in teleoperation and telepresence applications, among many others. Furthermore, motion control including retargeting and interaction are still open and difficult problems, that may benefit from the above-mentioned new representations. How to retarget the motion of the user on an avatar simulated with these new representations? In traditional rendering pipelines, how to benefit from new information, such as the surface of the user, to adapt the pose of the avatar in real-time? Alternatively, how to control this avatar, with all the possible representations, using only a few sensors (e.g. controllers, inertial sensors…)? All these new possibilities may have various impacts on the user experience. How can new representations, new control metaphors or devices, and new ways of doing character simulation, enhance the VR experience of users? How can we evaluate these new methods to measure the degree of interaction, embodiment, agency, and impact? What are the potential risks and limitations of these approaches from the user perspective? The goal of the workshop is to discuss different aspects of the most recent works in this domain, and explore how it would impact future research in VR. After scientific presentations, showcasing the latest research advances in avatar representation, simulation and control, a round table will help to imagine the next generation of avatars, and their impact on user experience.

We call for submissions of research papers and technical notes (4-8 pages) or position papers and work-in-progress research abstracts (2-3 pages) on the following topics:

§  Avatar modeling,

§  Avatar simulation,

§  Avatar control,

§  Avatar rendering,

§  User experience with avatars,

§  Interaction paradigms with avatars,

§  Ethics, GDPR and regulations related to avatar use

Accepted papers will be presented in scientific sessions, encouraging interaction with the audience (15 minutes talk + 15 minutes discussion) to help imagining the future trends of avatar modeling, representation, control and interaction. A round table composed of selected contributors will conclude the workshop: “Next Generation of Avatars: promises and limits”.

All the submissions should follow the IEEE Computer Society VGTC conference format (https://tc.computer.org/vgtc/publications/conference/).

Submissions must be camera-ready. NGA2025 uses double-blind reviewing, so please remove all personal data, such as author names and affiliation from all your submission files. 

 Papers should be submitted thanks to the Easychair system:

https://easychair.org/conferences/?conf=nga2025

Important dates:

§  Submission deadline: January 10, 2025

§  Notification: February 3, 2025

§  Camera-ready deadline : February 21, 2025

§  Workshop date : March 8-9, 2025

Organizers:

§  Franck Multon, Inria, franck.multon@inria.fr (contact)

§  Ludovic Hoyet, Inria, ludovic.hoyet@inria.fr

§  Quentin Avril, InterDigital, Quentin.Avril@InterDigital.com

§  Victoria Fernández Abrevaya, Max Planck Institute for Intelligent Systems, victoria.abrevaya@tuebingen.mpg.de