At the moment, there is a schism as to how to collect the data for gesture recognition in a general sense, and sign language recognition in particular. In general, the approaches are divided into one of two schools (although there also exist several hybrids):