Action annotation for video segments about vidat HOT 3 CLOSED

When labelling actions, after submitting an entry, have the end of the last action jump to be the start of the next action (saves a lot of time).

from vidat.

• Hierarchical action segments : assembly -> sub-assembly -> atomic actions. Allow user to first label sub-assemblies (assemble leg) and then separately load them and label atomic actions within each sub-assembly (pick up leg, screw-in leg, tighten leg). Only implemented action and objects because it is more generalize.
• Atomic actions: composed of drop-down menus of verbs and drop down of objects + option to add free description text. drop-down lists to be loaded from taxonomy files. should automatically filter out impossible pairs (can't screw in a table-top).
• Have the database allow for multiple action labels for the same frame (someone can spin a leg with one hand and pick up another leg with the other hand).

from vidat.

When labelling actions, after submitting an entry, have the end of the last action jump to be the start of the next action (saves a lot of time).
#54 (comment)

from vidat.