Scenario general view. Robot pursues the tracked person.
Scenario general view. Robot stares at each tracked person.
NAO-MPVS- ( number of people in the scenario )( control strategy ) - ( the sequence index )
The control strategies are defined as:
(e.g. 'NAO-MPVS-2F-1' for scenarios of 2 people, robot pursues the first tracked person)
Left image : robot left camera view
Blue bounding box represents face detections
Red bounding box represents tracking results
Green arrow represents the tracked person's self-motion projected on the image
Magenta arrow represents the tracked person's motion owing to camera motion
Right image : robot top view
Red circle represents each person's current postion and size
Blue line represents the person's trajectory
(NAO-MPVS-1-1) One moving person. Robot pursues him.
(NAO-MPVS-1-2) One moving person. Robot pursues him.
(NAO-MPVS-1-3) One moving person turn back. Robot pursues him while he is visible.
(NAO-MPVS-2F-1) 2 people and one go out of the scene. The out-image person is not moving. Robot pursues the first tracked person.
(NAO-MPVS-2F-2) 2 people and one go out of the scene. The out-image person is moving. Robot pursues the first tracked person.
(NAO-MPVS-2F-3) 2 people are moving with occlusions. Robot pursues the first tracked person when this person is visible.
(NAO-MPVS-2J-1) Robot stares at each person for 3s to maintain a large field of view.
(NAO-MPVS-2F-4)2 people turn back. Robot pursues the first tracked person when this person is visible.
(NAO-MPVS-3J-1) Robot stares at each person for 3s to maintain a large field of view.
(NAO-MPVS-3F-1) Robot pursues the first tracked person while tracking all people appeared in the image.