Abstract—The study is focused on the recognition of Logosyllabic Han writing system such as Cantonese for Chinese, Hiragana and Katakana for Japanese and Hangul for Korean from a camera-captured image using image processing algorithms: Peripheral Direction Contributivity and 3-level matching, 2D Array and Recursion methods. In order to improve performance and to reduce memory usage of a mobile device, a remote web server has been used to communicate information from the user to the server and vice versa using Simple Object Access Protocol. Experiments were conducted to test the accuracy of the application such as capturing the same image several times, checking the percentage of the features of the text in the captured image, and capturing text from images in different circumstances such as low lighting and skewed images. The overall result showed the application has 80-100% accuracy in identifying Chinese, Japanese, and Korean writing systems.

Index Terms—Image processing, logosyllabic han character, optical character recognition, peripheral direction contributivity.

The authors are with Technological Institute of the Philippines, Quezon City, Philippines (e-mail: christina.aragon@tip.edu.ph, melissa_tipqc@yahoo.com, rjmcabauatan@ yahoo.com).

Cite:Maria Christina Aragon, Melissa Juanillo, and Rosmina Joy Cabauatan, "Camera-Captured Writing System Recognition of Logosyllabic Han Character," International Journal of Computer and Communication Engineering vol. 3, no. 3, pp. 166-171, 2014.