HeadCount technology employs a patented method of counting with artificial intelligence by means of an advanced pattern recognition engine.
Every HeadCount counting device contains an onboard neural network. A neural network is an emulation of the human brain in an electronic format.
Thus, if a person walks through the count zone the pattern generated by the sensors is sampled and then processed by the onboard neural network that has been trained to recognize the number of people passing by and in which direction.
A stimulus (in HeadCount’s case the pattern generated from the counter) is presented to the network which compares it to learned events, which when recognized, outputs a signal (in this case accurate counts of ‘in’ and ‘out’).
Training of the neural network involved “show and tell” conditioning of thousands of patterns which were recorded in various locations creating a representative a priori data set. The output result is obtained by neural calculation for each pattern presented.
Click here to view Counting Technology Animation
|
