We are proud to announce that LIBRA AI Technologies has received Springer Best Paper Award at the 20th International Symposium on Visual Computing (ISVC 2025), held in Las Vegas, USA.

Our paper, “Gaze Prediction in Virtual Reality Without Eye Tracking Using Visual and Head Motion Cues”, was recognised for its scientific excellence, technological innovation, and practical relevance in the fields of Virtual Reality (VR) and Visual Computing.

ISVC is one of the leading international forums for research in computer vision, VR, visual analytics, and AI. Its proceedings are published by Springer in the prestigious Lecture Notes in Computer Science (LNCS) series, making this distinction an important milestone for our team and a strong validation of our R&D capabilities.

Turning Research into Practical VR Innovation

Eye tracking is widely considered a core component of immersive VR systems. It enables capabilities such as foveated rendering, adaptive scene optimisation, and attention-aware interaction. However, in commercial environments, eye-tracking hardware often presents challenges:

  • It is not available in most headsets
  • It increases device cost and integration complexity
  • It may introduce privacy and data governance concerns

Our research addresses this gap directly.

Instead of relying on specialised sensors, we developed a multimodal AI framework that predicts user gaze using existing VR signals. By combining:

  • Head motion data from VR devices,
  • Visual saliency cues extracted from the scene,
  • Advanced temporal prediction architectures (LSTM and TSMixer),

we can forecast user gaze direction up to 333 milliseconds ahead, outperforming standard baselines such as Centre-of-HMD and Mean Gaze.

The result is a solution that is:

  • Real-time capable
  • Lightweight and computationally efficient
  • Deployable across Linux and Apple ecosystems
  • Privacy-preserving by design

This approach removes a key hardware dependency and makes advanced gaze-aware functionality accessible to a broader market.

Why This Innovation Matters

For VR platform providers, content creators, and technology integrators, this innovation unlocks tangible advantages:

  • Enable foveated rendering without built-in eye tracking
  • Reduce computational load through predictive focus adaptation
  • Improve user immersion by minimising perceptual latency
  • Build scalable, privacy-compliant VR applications
  • Lower hardware requirements and associated costs

In practical terms, this means organisations can deliver high-quality immersive experiences without investing in specialised hardware infrastructure.

For sectors such as digital entertainment, live-streamed events, training simulations, cultural experiences, and enterprise collaboration platforms, predictive gaze modelling creates new opportunities for adaptive, attention-driven interaction.

Bridging Scientific Excellence and Real-World Impact

The Best Paper Award at ISVC 2025 recognises not only technical contribution but also our ability to convert advanced AI research into deployable solutions.

As a boutique data science agency, LIBRA AI Technologies bridges research innovation and operational implementation. From precision agriculture with Edge AI to immersive VR systems, we focus on delivering measurable value through robust, accessible AI technologies.

This distinction further strengthens our position as a trusted R&D and innovation partner across both European research initiatives and commercial technology development.

Acknowledgement

This work was developed within the framework of ACCESS – AI and VR-based technologies for live-streamed performing arts (HORIZON-MSCA-2022-SE-01).

We sincerely thank our collaborators and co-authors from the ACCESS consortium for their partnership and support.