摘要 |
PURPOSE:To attain a high speed and to display and discriminate automatically, by collating again candidate characters corresponding to characters to be corrected by a dictionary for correction where much standard characteristics are stored than that of the content of a recognizing dictionary. CONSTITUTION:An input data 1 is converted into binary information at an observing section 2. The characteristic of the input character is extracted from this information at a characteristic extraction circuit 3. This characteristic is stored in an input character characteristic storage section 4 and collated with a recognizing dictionary 5 at a collation circuit 6. Most useful and a few standard characteristics only to each of the character number to be collated are stored in the dictionary 5. As the result of the collation, the candidate characters from the 1st to the K-th rank are stored in a result of collation storage section 7 and the result of the 1st rank is displayed on display 8. The character to be corrected is designated from a correcting character indicating section 9 by observing display. Further, the characteristic of the inputted character is collated 6 by reading out a lot of standard characteristic number from the dictionary 10 for correction, and the closest candidate character is displayed automatically on the display 8 as the result of recognition. |