Digit recognition system using EEG signal

Abstract

Based on Linear Discriminate analysis (LDA), Principle Component Analysis we explore the characteristics of multichannel Electroencephalogram (EEG), which is recorded from no of subjects recognizing different numbers displayed on the screen by a GUI software designed in Visual Basic 6. The scaling exponent of each digit is different especially at positions C3 and C4, and at positions O1 and O2. LDA exhibits its robustness against noises in our works. We could benefit more from the results of this paper in designing mental tasks and selecting brain areas in brain-computer interface (BCI) systems. The objective of this system is to report the work done related to sensitivity of EEG signals related to specific thought process. The thought process was chosen to be numbers (0-9). The main objective of this work is the analysis and classification of EEG signals among the men and machines and provide a secure communication interface. EEG recordings of six male right-handed subjects in the age group of (20-25) were taken. The subjects were normal without any mental disorder. They did not have any problem in communicating and had normal vision. All subjects have good knowledge of digits. A simple display system in visual basic is prepared for the project. This system generates random number with interval of 2 s. After every 2 s a random number is displayed on the screen. The recording was captured for 3 min. This process was repeated for five times. The EEG signal has been processed by statistical analysis methods such as LDA and PCA. It was found that the EEG signals are sensitive to thought process. So it is possible to recognize thought process through EEG signals. In our ten digit thought process, we get ten distinct clusters by analyzing EEG signals through statistical technique like LDA and PCA. The recognition rate of LDA is 70%. The recognition rate of PCA is 37%. So we establish that LDA is more powerful method as compared to PCA. We observe that the EEG signal is more dominant on right hemisphere as compared to left hemisphere. The data base created has potential to be used as a digital recognition system. It has tremendous applications in design of security system.