抄録
Three different levels of attention is realized: broad attention to whole environment, detail attention for the part of the focused human, and realtime attention for the 3D motion of the whole body of the focused human. We developed the humanoid head with wide viewing angle and high visual acuity by using five cameras and Image Recognition Module. RT-Middleware enables several simple functions to work together very easily. With the systems, the humanoid can observe human attention efficiently.
