A PATTERN RECOGNITION SYSTEM FOR A NEW LASER SENSOR

This paper presents the concepts that are behind of a new real-time, non-contact, static hand sign recognition system. This system is a new embodiment of a human computer interface used to determine the emotional state through the body movements and postures. This new application shows the versatility of the concepts on which both systems presented above relies on. The new proposed recognition system is an independent embedded intelligent platform able to classify different hand signs through a neuronal network. The system works on a DSP device that commands the laser scanner system (a component of the new transducer used to extract hand sign characteristics), acquires images, extracts hand features, preprocesses the features, classifies them and, at the end, vocalizes correspondingly each recognized hand sign.