To answer your first question, we had both bounding boxes added and removed, and depending on the dataset, the main type of error was different (I'd say it was

To answer your first question, we had both bounding boxes added and removed, and depending on the dataset, the main type of error was different (I'd say it was overall more objectifs that were forgotten, especially small objects). It would indeed be very interesting to see the impact of those improved datasets on driving, which is ultimately the task that is automated for cars. We've been working on many projects at Deepomatic not only related to autonomous cars, and we did see some concrete impact of cleaning the datasets beyond performance metrics.