In the first scenario, self-resentation, the subject is recorded while briefly introducing himself in front of a webcam mounted on a desktop monitor. The dataset contains 89 videos, with an average length of 1min30sec.
Each subject filled in the Italian version of the Big Five personality questionnaire. For more details see (Batrinca et al, 2011a) and (Batrinca et al, 2011b).
In the second scenario, Human-Human Interaction, the subject (Instruction Giver) sat in front of a computer monitor, on which a map was displayed. His/her task was to guide another person (confederate – Instruction Follower), in another room, by following the route marked on the map. The interaction was captured by a webcam mounted on the monitor and a lapel microphone. Four different behaviors were displayed by the confederate, which ranged from fully collaborative to fully non-collaborative. The dataset contains 35 videos, with an average length of approximately 4min/behavior. Each subject filled in the Italian version of the Big Five personality questionnaire. For more details see (Batrinca et al, 2012).
In the third scenario, Human-Machine Interaction, the subject was interacting with a computer, and was told to guide it from the start point to the end point, according to the route marked on her/his map. Each subject filled in the Italian version of the Big Five personality questionnaire. The third scenario is almost identical to the second scenario. The only difference being that the subject was interacting with a computer (WoZ approach, where a human operator is simulating a robot with an artificially sounding voice). The dataset contains 54 videos, with an average length of aprox. 5min/behavior. The paper describing this dataset is currently under review. Please stay tuned or contact us for details.
