Autonomous Vehicles

AvidBeam built a tool for manual and automatic annotation for video and LiDAR. Upcoming phases will see the generation of perception algorithms for different scenarios.

Flexible architecture with a diversified set of features to ease automation of the complex-annotation process

Diversified annotation scenarios with built-in intelligence to detect, recognize and relate objects

  • Cloud-Based Architecture tested with major cloud suppliers (Amazon, Microsoft Azure, IBM Soft-layer)
  • Automatic annotation for uploaded/stored data
  • 2D/3D Annotation mapping for Video & Point Cloud
  • Integrate Drive testing metadata with annotation results (Country, Location, Driver Name, GPS…)
  • Built-in Manual annotation support to review and edit annotated data
  • Functional annotation support
  • Curb detection
  • Scalable Open platform with open API’s to add new CV/AI algorithms easily
  • Ability to retrain built-in algorithms or train new CV/AI models for better recognition accuracy, and to leverage system automation
  • Flexible and Easy integration with content management systems and ADAS testing system
  • Semi-Automatic tracking for manually annotated objects
  • Elastic search for annotated data
  • Scene management and scene search
  • Object tagging with re-sizable bounding box
  • 3d point cloud clustering