Analyzing elektrooculogram signal (EOG) and modelling by the artifical inteligience methods

Biomedical technology is a science which has subject matter to develop all instrument for diagnose and treatment of disease. Biomedical equipment is common research topic in engineering and medical science. Owing to fact that the benefits of the biomedical technology increase, significance of this instrument is increasing in the same line. New biomedical equipment have been tried to produce for paralyzed person to raise their life quality. At present day, affected movements have been obtained by utilizing other organs for paralyzed patient. So eye movements have become important data source. Especially the usage of eye movements for giving message to outside is popular scientific subject. In studies according to eye movements, the electrooculogram (EOG) signal is used. EOG signals, influenced on age, gender, lighting, can be analyzed with different methods. In this thesis study, an intelligent control system which can detect direction with EOG signal has been referred. The vertical and horizontal EOG signals taken from electrodes, placed around the eyes, and has been modeled by using Artificial Neural Networks and Fuzz Logic which are artificial intelligent techniques. The system can sense four main directions (Right, Left, Up and Down) at the same time it can detect blinking and tic movements. Firstly, the signals have been cleaned from noises and needless parts by amplifying and pre-filtering. The "movement range" and "movement switching control" algorithms have been recommended for feature extraction from EOG. The eye movements can be perceived as vertical and horizontal with these algorithms. According to these features obtained signal can be classified with intelligent control systems. The performances of the recommended ANN and Fuzzy logic control models have been demonstrated by analyzing the statistical accuracy. It has seen that each model can be classified the eye movement successfully; however it is observed that the fuzzy logic model has better performance than the ANN model. The user interface application has been developed to follow eye movements from computer screen. Direction data and EOG features of patients have been projected on the screen with this application. Finally, in this work, a fuzzy logic control model has been designed to detect EOG direction of people which have diplopia. It is shown that the eye movements have been determined successfully by using diplopia data and EOG features in test results.