Novel methodology for blink recognition using video oculography for communicating

Communicating by those suffering from motor neuron disease has always been an arduous task. A lot of research has been carried out in finding new methods to assist this section of the society in order for them to freely communicate with the outside world. Though several methodologies exist using electrooculography or video oculography, most methods involve complex algorithms which are slow, power hungry and also area inefficient. In this paper we introduce a new proposition, that simple eye blinks could be used by these individuals in order to communicate. These blinks further on could be converted to Morse code to transmit messages, where in each blink represents a dot or a dash. In order to achieve the same, a novel algorithm to first identify the ocular region of the face and also to further identify the blinks using image processing has also been presented. All algorithms were designed and developed for a Spartan 3e series of FPGA and were tested using MATLAB 2011b software. Experiments were conducted under various lighting conditions for a wide set of people and an accuracy of 92% to identify different blinks was achieved.