Word string collating apparatus, word string collating method and address recognition apparatus

A word string collating apparatus for collating an input word string and words in a word dictionary when a partial word string is extracted from the result of character recognition for a word string including a plurality of words, comprising:

• word string forming means for forming a word string based on the result of character recognition for a word string including a plurality of words;
• correspondence setting means for variously setting correspondence relations between the words of the word string formed by said word string forming means and the words in the word dictionary according to the number of words of the extracted partial word string;
• distance calculating means for deriving each distance between the words which are set into the correspondence relation by said correspondence setting means based on each similarity between the words;
• positional relation deriving means for deriving a positional relation of each word of the formed word string which is set into the correspondence relation by said correspondence setting means;
• evaluated value deriving means for deriving an evaluated value based on the positional relation derived by said positional relation deriving means and the distance, derived by said distance calculating means, between the words which are set in the correspondence relation by said correspondence setting means for each type of the correspondence relation set by said correspondence setting means;
• and
• determining means for determining a partial word string extracted from the formed word string based on the evaluated value derived by said evaluated value deriving means.

A word string collating method for collating an input word string and words in a word dictionary when a partial word string is extracted from the result of character recognition for a word string including a plurality of words, comprising the steps of:

• variously setting correspondence relations between the words of the input word string and the words in the word dictionary according to the number of words of the extracted partial word string;
• deriving each distance between the words which are set into the correspondence relation in said correspondence setting step based on each similarity between the words and deriving the positional relation of each word of the input word string which is set into the correspondence relation in said correspondence setting step;
• and
• deriving an evaluated value based on the positional relation derived in said deriving step and the distance between the words which are set in the correspondence relation in said correspondence setting step for each type of the correspondence relation set in said correspondence setting step and determining a partial word string extracted from the input word string based on the evaluated value.

A word string collating apparatus for collating an input word string and words in a word dictionary when a partial word string is extracted from the result of character recognition for a word string including a plurality of words, comprising:

• forming a word string based on the result of character recognition for a word string including a plurality of words;
• variously setting correspondence relations between the words of the word string formed in said word string forming step and the words in the word dictionary according to the number of words of the extracted partial word string;
• deriving each distance between the words which are set in the correspondence relation in said correspondence setting step based on each similarity between the words;
• deriving a positional relation of each word of the formed word string which is set into the correspondence relation in said correspondence setting step;
• deriving an evaluated value based on the positional relation derived in said positional relation deriving step and the distance, derived in said distance calculating step, between the words which are set in the correspondence relation in said correspondence setting step for each type of the correspondence relation set in said correspondence setting step;
• and
• determining a partial word string extracted from the formed word string based on the evaluated value derived in said evaluated value deriving step.

A word string collating apparatus for collating words of an input first word string including a plurality of words and words of each of third various word strings of a word dictionary when a second word string using part of the plurality of words is extracted from the result of character recognition for the first word string, comprising:

• character recognizing means for recognizing the first word string containing the extracted second word string in the unit of character;
• word extracting means for extracting characters recognized by said character recognizing means in the unit of word;
• and
• word string extracting means for collating the first word string including a plurality of words extracted by said word extracting means and the third various word strings of the word string dictionary, determining words of the second word string in the first word string respectively corresponding to the words of the third word string based on similarities between the words of the first word string and the words of the third word string, making evaluation for each of the third word strings based on the number of words between the words in the second word string thus determined and the similarities between the words of the third word string and the words of the second word string determined, and extracting one of the third word strings as the second word string.