Development of computer based eye-writing system by using EOG

There are various recommended methods for patients with Amyotrophic Lateral Sclerosis (ALS) and Locked-in Syndrome (LIS) as well as for paralyzed patients who are not able to speak and/or not able use sign language for communication. In very few cases, the eye muscles become paralyzed for these patients. Thus, based on this evidence, the methods that code eye movements into characters used for writing surpass the others. The speed of existing methods that code eye movements into characters by using camera or the electrooculogram (EOG) is about 10 words per minute in general. In this study, firstly, the methods developed for coding eye movements into characters were analyzed in order to determine coding speed and the accuracy rate of the character that is aimed by an eye movement in these techniques. Based on the results of this study, a new eye movements coding technique, which depends on the eye movements in 4 different directions and 2 different view angles and different waiting time in each of these directions with characters coded based on in each specific eye movement direction, is suggested. Computer based system is developed for implementation of suggested technique. This developed technique consists of two parts; hardware and software. While hardware part records EOG signals that occur during horizontal and vertical eye movements and puts this information into a computer, software part infers and vocalizes these signals by coding these into characters consisting letters, numbers and punctuation marks in order to have proper words and sentences with the help of programs. The efficiency of this newly suggested technique is analyzed through assessing parameters that indicate the speed of coding eye movements into characters and the accuracy rate of coding. The speed of coding eye movements into characters with suggested new technique developed in this study is 17.5 words per minute. While the accuracy rate of coding into right characters in the first trial of a text comprised of 126 letters is 92.5%, this accuracy rate reaches to 100% starting from the fourth trial of the technique. Moreover, this repetition of using this new technique improves both the speed and the accuracy rate of coding into right characters. All these findings imply that even though this newly developed methodology requires a settling-in period needed for grasping how to code efficiently eye movements into characters, after passing this period, this new eye movement coding method developed in this study would be able to improve the speed of coding at least 50% faster than the previously suggested similar ones.