PURPOSE: To improve recognition accuracy by recording a word constituted by means of a character which comes to be the object of character recognition and frequency data addopted after the word is character-recognized in a word dictionary by a time group.
CONSTITUTION: A character recognition processing part 1, a word generation part 2, the word dictionary file 3, a word selection processing part 4, a word decision processing part 5 and a display device 51 are provided. The word constituted by the character which comes to be the object of character recognition and frequency data addopted after the character recognition of the word are recorded in the word dictionary file 3 by the time group and the word selection processing part 4 picks up frequency data held by a candidate word from the word dictionary file 3 so as to generate a selection word data part 31. Therefore, character selection with high character recognition accuracy considering frequency data by the word decision processing part 5 is executed.
FUJISHIMA YOSHIHISA