[Japanese | Thesis | Researches in Minoh Lab | Minoh Lab]
We propose an automatic imaging method of speaker's situation for distance learning using various sensors.
The method consists of two processes: an estimation of speaker's situation, and controls of cameras and a selection of video images. We use multiple cameras, a microphone array and a position sensor in order to detect information such as the position and the direction of the speaker. By fusing same kind of information detected from various sensors, we can improve a precision, expand a region and increase opportunities of detecting. By fusing different kind of information, we estimate a speaker's situation.
We applied our method to simulated lectures. As the result of the questionnaire survey, it is shown that our method is effective in imaging of speaker's situation.