In a feature published in Kathimerini, one of Greece’s most established and widely read newspapers, journalist Sofia Christou discussed with LIBRA AI Technologies CEO Yannis Kopsinis about the company’s work on a real-time system for detecting driver fatigue and distraction. The system is part of LIBRA’s broader strategy to develop advanced, tailor-made AI solutions from scratch, addressing real-world safety challenges through state-of-the-art technologies.
One of the system’s core technologies is eye tracking, and development began with a research collaboration between LIBRA and the I-SENSE group of ICCS/NTUA, focusing on building and testing an early prototype in real driving conditions. Building on that foundation, LIBRA went on to design and implement a pilot version of a drowsiness detection subsystem, which was deployed and evaluated in operational conditions at the Piraeus Container Terminal (PCT). These early pilots offered valuable insights and shaped the direction for further development.
The system has since evolved into a multi-parametric AI framework, capable of analysing facial expressions, blink rate, yawning, gaze direction, and emotional state — all processed in real time on edge devices. A key innovation lies in the system’s ability to assess not only whether the eyes are open, but also where and what the driver is focusing on, offering a more accurate picture of attentiveness. True to LIBRA’s principles, the system is designed to operate without transmitting or storing raw visual data, using non-reversible embeddings to ensure full GDPR compliance.
Development continues today, progressing toward production with LIBRA investing its own resources, but also as part of LIBRA’s technology contribution to the European HORIZON H2020 project HIDDEN, in collaboration with CIBOS, a spin-off of ICCS. Within this initiative, LIBRA serves as the Technical Manager in a consortium of 14 organisations, coordinated by ICCS. The technology is currently being tested in semi-autonomous vehicles, under demanding, high-risk urban scenarios, where timely detection of driver inattention is critical for safety.
Beyond private vehicles, the potential applications of this technology extend to buses, trains, industrial machinery, control rooms, and remote operations — any environment where human focus plays a central role in operational safety. In such professional settings, AI can act as a support layer, detecting early signs of fatigue or distraction before they turn into risks.
In addition, individual subsystems, such as camera-based gaze tracking, can be adapted for independent use in other sectors. One example is in retail, where it can help analyse consumer attention and preferences at the shelf, offering valuable insights into behavioural patterns.
At LIBRA AI Technologies, we are committed to building specialised, cross-sector AI solutions. We believe that adaptability and interoperability are key to unlocking the full potential of artificial intelligence — not only to enhance safety, but to better understand and support human behaviour across environments.