| 摘要 |
PURPOSE:To obtain just a single writing way of addresses to be registered in an address dictionary contained in an address reader by detecting a key word during a process when the word strings forming written addresses are extracted for place names from a major section to a minor section. CONSTITUTION:The character read results are outputted in series from a character reading part 10 for each line from the left end to the right end and then set to a word buffer 23 via a word extracting part 21 and a digit code extracting part 22. The information on word strings is inputted to an address component word extracting part 24, and an address is decided in the order of the major, medium and minor sections of place names. This decided address is supplied to an input word selection part 28 via the buffer 23. A maximum coincidence detecting part 31 collates the word sent from the part 28 with a key word set at a collation dictionary buffer 43. Then no writing conversion is carried out if no coincidence is obtained from said collation, and the collated word is set to a writing conversion word buffer 33. A place name deciding part 52 selects some candidates out of the maximum coincidence detected place names in the order of higher degrees of coincidence. These candidates are delivered to an address deciding part 71 and outputted after integrating the writing ways.
|
