A writing system for brush characters using neural recognition and fuzzy interpretation

Presents a fuzzy-based writing system for brush characters. The system converts one-stroke-at-a-time mouse input to brush characters. To recognize an input stroke, the system uses a neural network. Accordingly the individuality of user-input characters can be reflected in each stroke of the character by this feature of the system. The individuality in this case is the speed of the moving mouse, the length of the stroke and the position of the stroke. From the speed of the moving mouse, the system fuzzily controls the change of the output stroke shape with the speed parameter. As a result, the output characters have a natural scratched look and length depending on the input stroke.