Analysis of vision-based Text Entry using morse code generated by tongue gestures

We propose a Text Entry Interface based on the detection of tongue protrusion gestures using Computer Vision methods. The system uses a common webcam to acknowledge gestures made with the tongue, which are interpreted as the “dots” and “dashes” of a double-switch Morse Code. One difficulty of employing an encoding technique for text entry is the necessity of memorizing the code in advance. Thus, to enable the rapid utilization of the system, the user is provided with a on-screen chart that provides immediate visual feedback. The 30-character chart has been designed as a dichotomic search table that matches perceptually the position of the tongue gestures. A theoretical speed of 1.44 WPM is calculated using moderate parameters; however, experimental results exceeded the theoretical speed by 50%. Although our interface has been tested with only two users, we demonstrate that the performance increases as the level of expertise of the user augments.