Recognition of Hand Writing Japanese Character

Handwriting is one of the most natural writing ways of characters. A Japanese brush pen is one of traditional writing material, and it is used in practice of the writing characters in schools. We developed a new computer input device using a brush pen. The written character is reconstructed by contour sequences of the tip of the brush pen by using image processing technologies. The written characters are classified by two pattern recognition methods. One is based on the use of the chain codes for each brush pen stroke. And the other is based on the use of the marginal distributions of the reconstructed characters. The accuracy of the recognition for the characters is enough high in character recognition. Sampling time was 0.1 second. As the line width of character can be changed just as wanted, the developed computer input device is comfortable in writing the characters.