Meeting state estimation using head movement

This paper discusses a meeting state estimation method that uses head movement. To obtain head movement, face position movement is detected from images captured by the participants' cameras used in a multimedia meeting; no additional user equipment for meeting state estimation is needed. We extract head movement from detected face movement, and cross reference the movement and meeting state obtained from the results of a questionnaire completed by subjects who participated in the actual meeting. Finally, we estimate the meeting state of another actual meeting archive.