Data tools
Automatic annotation
Multi sensor AI-based automatic annotation for dynamic and static objects with industry-leading precision
Increases the proportion of auto-annotated sensor data to almost 100%
Removes the bias from existing pipelines
aiNotate can process the data close to real-time
Scalable for large fleet use cases
Automated edge-case mining and active learning to reduce storage and computation costs
Automated annotation and knowledge transfer to reduce annotation costs
Synthetic training data generation
The core of aiData is our high-fidelity sensor simulation, generated by aiSim, and enhanced by an AI-based reality filter, which can achieve outputs indistinguishable from the real world. This is extended by cinema-grade, high-fidelity renderings and augmented reality
Capable of modelling costly, dangerous, or rare sensor data in simulation
Makes the data indistinguishable from real world data using reality filters or cinematic CGI
Builds an automated pipeline for randomizing and large-scale data generation
Scenario extraction based on test drives and aiNotate
Automatically trains, tests and benchmarks the data quality; finds gaps to fill and improve