Teaching the computer to see...

in any situation

We focus on the creation of a robust system that does not only handle best visibility conditions but also solves worst case scenarios. Having also the computational limits for an embedded application in mind, we master the trade-off between speed and accuracy.
Below are some more real-life examples of our perception system.
Performance of our current system in selected scenarios for Autonomous Driving

* all videos are generated by the aiVis Freespace Detection in real-time

  • Nighttime highway driving
  • Low lighting
  • Construction zones
  • Strong shadows on road
  • No clear lane borders
  • Heavy rain highway driving
  • Distraction by raindrops & windshield wiper
  • Challenging to distinguish between rails and lane border