Image Morse code text input system

Many image processing techniques [1][2][3][8] and fuzzy algorithms [7][9][10] are applied in this research, such as color extraction, advanced morphology, particle filter, edge detection, membership function, fuzzification and defuzzification etc. The purpose of this paper is to show an image Morse code text input system based on real time mouth image recognition that provides the useful assistive communication interface for the disabled [3]. A web camera with USB interface is used in this system to acquire dynamic mouth images and then the image recognition will extract the feature of open/close mouth status. The time period of open/close mouth status can be counted and regarded as tone space/silent space of Morse codes which can be recognized by fuzzy algorithms we designed. After that the subject can type characters on the computer's screen by just opening/closing his/her mouth in front of the web camera achieving the functions of word processing. The experiment results indicate high accuracy can reach almost 95% if the subject accepts a period of training in advance.