| 摘要 |
PURPOSE:To improve a reading accuracy and to decrease a memory capacity and a processing time by collating a recognizing candidate character with a word dictionary outputted from a character recognizing means, detecting the place where the word including a recognizing candidate character to two continuous characters is absent and providing the border of a word string between them. CONSTITUTION:After a recognizing candidate character 11 to an input Japanese sentence 10 is obtained by a word recognizing means 3, a word collating means 5 executes the collation with a word dictionary 4. Before the word to start from the recognizing candidate character to the first character of the input Japanese sentence 10 is selected and the word to start from the recognizing candidate character to the second character is selected, a word string boarder detecting means 6 investigates whether or not the division between the first character and the second character comes to be the word string border. Namely, it is investigated whether or not the word including the recognizing candidate character or not the word including the recognizing candidate character to the first character and the second character is selected, and when such a word is not selected, the location is made into the word string border. Thus, the processing scope can be correctly and shortly set and the memory capacity and the processing time necessary to the processing can be decreased.
|
