The monitoring space registration function conducts bed and bedside spatial monitoring automatically by
determining, and then entering, the horizontal plane and corner locations of the bed into the system. Using this
function, the bed’s position in the measurement area can be determined and positional relations of a patient based on
the registered bed corner can be ascertained. The head detection function, which is based on depth information,
permits the patient’s head to be detected using cascaded classifiers based on Ada Boost and local binary pattern
(LBP) features[7,8]. These features use changes in the illumination pattern within the measurement area to detect
patient movements without significant difficulties.