In the AVData Webinar we spoke about Computer Vision training data labeling when looking at the Berkeley and Apollo dataset.

Here is a summary of 5 ways of modeling computer vision from the AIX - Designing AI Experience course. This is applicable to Autonomous Vehicles (and of course in other Computer Vision use cases in other industries)

https://driverlessworldschool.teachable.com/admin/courses/531002/curriculum/lectures/14444944