Abstract:
As the technology advances, touchscreen device became the most popular at the present time. This work proposes the prediction model for braille input and translation of braille characters to text in Thai for visually impaired with touch screen device. The model is needed for people with visual disability to improve accuracy braille sequence of touchscreen device and this paper proposes method that compensate for finger drift and this method has the effect of following a users touches over time. With FDC model, coordinates are input into a device with multi-touch, based on the six dots of the braille code of brailler, which each set of coordinates represents one character. Sets of coordinates are grouped based on their timings so that the coordinates in each group are hypothesized as belonging to the same character. This paper is designed such that it learns the user's typing behavior including individual typing rhythm and drifting patterns in order to translate the inputted braille keystrokes to the corresponding Thai characters. The experimental result from four blind participants showed 89.23% accuracy when using the proposed method.