Recognition of Six basic facial expression and their strength by neural network

Develops an `Active human interface´ that realizes interactive communication between machine (computer and/or robot) and human. The authors investigate the method of machine recognition of human facial expressions and their strength. They deal with the neural network method of recognition of facial expressions. Considering 6 groups of facial expressions, i.e. surprise, fear, disgust, anger, happiness and sadness, they obtain 30 x- and y-coordinates of facial characteristic points representing 3 face components (eyes, eyebrows and mouth). Then they generate the facial position information which is input to the input units of a neural network; the network learning is done by backpropagation algorithm and the recognition test is carried out. For the six basic facial expressions, the correct recognition ratio was found to be about 90%. This paper further investigates the method of recognizing the strength of the six basic facial expressions by a neural network