Arabic Code Generation based on Four Direction of Human Eye

Abstract

In this paper, we built a Desktop Application for a new encoding method presented to help disabled Arabic-speaking people. The proposed encoding method depends on the direction of the pupil of the eye so that the most frequently used letters have the shortest codes. The language characters are converted into a sign of the direction of movement of the human eye's pupil. The proposed system is a text entry system consisting of three parts. The first part included a CNN-based algorithm for evaluating the direction of the eye pupil, and the second part was an algorithm for building an encoding triple tree to generate the symbol for Arabic letters, while the third part involved Translate codes to its corresponding letter. The system was tested by a number of well-trained participants. The accuracy of the system goes up to 99.3% while the system's sensitivity reached 88.5% and total specificity 99.6%. also SCR and MSD metrics are computed.