I thank its structure consist of 3 parts: Hardware: input instrument include major video camera, also of keyboard etc. data USB interface. Output: monitor and loudhailer. Software: be composing initialize definition system (corresponding initialize things name and the words sound, natural digital values); operate definition recognition by oneself, namely machine learning by itself; and definition program of combine pattern by self (make the computer having creativity ). The core problem is extract and recognize graph from the pictures. If only find the Natural ratio of object of between initialization similar values. Other: Some Inspection equipment and coordinated software.

What's the actual question?