| 摘要 |
PURPOSE:To improve a recognition speed by composing standard patterns of respective characters in word units, collating the standard patterns with partial patterns in a character line image and finding similarity and terminals of the partial patterns, and determining starting terminals to be collated with the standard pattern of a next character. CONSTITUTION:A similarity detection part 2 inputs plural standard patterns constituting a word name to be collated from a standard pattern storage part 4 successively through a word table 3 and further inputs the starting terminal of a character line image to be collated from a starting germinal detection part 5 and then the similarity with a coincident partial pattern and the ending terminals of the partial patterns are outputted. The similarity is added to the last accumulated similarity by an addition part 7 and the result is stored in an accumulative similarity storage part 8. The ending terminals of the partial patterns, on the other hand, are transferred to a starting terminal detection part 5, which finds the starting terminal of a next partial pattern to be collated with the standard patterns constituting the word name and it is transferred to the similarity detection part 2.
|
