We ubiquitously placed stereo cameras around a room and connected them to computers via a large-capacity gigabit network. We placed the stereo cameras statically, utilizing range information as a key, and the system could successfully recognize human poses and gestures in the scene.

    Ubiquitous Stereo Vision (USV), which we propose here, is defined as having features listed in Table.

    The objectives of this research are to use 3D and color information, and to develop a real-time new human sensing method.

    Scene division by 3D information.

    Investigates human centered descriptions.

    Obtains personal log with personal identification.