How many tasks to generate for the case that there are several sensors on one subject

[RobvdDuim]

You could choose to make a task for each sensor_A-sensor_B overlap. This would result in a video having to be annotated more than once if a subject has several sensors on it at the time.

In the current design the sensor_A-sensor_B combination for a given subject with the longest overlap gets chosen for task generation.