An EOG-based virtual keyboard for paralyzed/physically disabled people

Abstract

There are many severely paralysed patients around the world and they are only able to move their eyes. In order for them to communicate, a system will need to be developed to make use of the eye movements to generate a message. To detect eye movements, electrooculograms (EOG) was chosen. In this project, a system comprising of a virtual keyboard based on EOG signals has been developed. There are many implementations of virtual keyboards that are done before using EOG signals. However, this research field is relatively new and new design and implementation can be done to improve the performance. Threshold values to detect various EOG signals were used in this proposed system. For the virtual keyboard design, multi-dimensional grids were created to optimize the typing speed using only four-directional eye movements. From the experimental results, a typing speed of about 27.69 characters per minute is achieved with 93.48% accuracy, which proves that the proposed system has potential use in practical situations.