OCR error correction
AssigneeFirst American Financial Corporation
InventorJeffrey Norton
Life of this patent
Estimating the term…
Abstract
Implementations of the disclosure are directed to OCR error correction systems and methods. In some implementations, a method comprises: obtaining, at a computing device, optical character recognition (OCR) text extracted from a document image, the text comprising a token; searching, at the computing device, based on a token bigram determined from the token and a mapping between words in a corpus and a corpus bigram set comprised of unique bigrams from the beginning or ending of the words in the corpus, the corpus for a best word to replace the token; and replacing, at the computing device, the token with the best word.
Claims17
A non-transitory computer-readable storage medium storing instructions that, when executed by a processor, cause a system to perform operations comprising:
- obtaining optical character recognition (OCR) text extracted from a document image, the OCR text comprising a token;
- searching a corpus for a best word to replace the token by performing operations comprising:
- determining one or more homoglyph pairs of each of a first token part and a second token part of the token, or each of a last token part and a penultimate token part of the token;
- determining a token bigram as one of the one or more homoglyph pairs;
- and
- using a mapping that groups words in the corpus based on unique bigrams from the beginning or ending of the words in the corpus;
- and
- replacing the token with the best word.
A non-transitory computer-readable storage medium storing instructions that, when executed by a processor, cause a system to perform operations comprising:
- obtaining optical character recognition (OCR) text extracted from a document image, the OCR text comprising a token;
- searching a corpus for a best word to replace the token by performing operations comprising:
- determining a first set of one or more homoglyphs from a first token part of the token or a last token part of the token;
- creating a first list of one or more homoglyphs from the first set;
- determining a second set of one or more homoglyphs from a second token part of the token or a penultimate token part of the token;
- appending or prepending each of the homoglyphs of the second set to at least one of the homoglyphs of the first list to create an updated first list;
- and
- using a mapping that groups words in the corpus based on unique bigrams from the beginning or ending of the words in the corpus;
- and
- replacing the token with the best word.
A method, comprising:
- obtaining, at a computing device, optical character recognition (OCR) text extracted from a document image, the text comprising a token;
- searching, at the computing device, a corpus for a best word to replace the token by performing operations comprising:
- determining one or more homoglyph pairs of each of a first token part and a second token part of the token, or each of a last token part and a penultimate token part of the token;
- determining a token bigram as one of the one or more homoglyph pairs;
- and
- using a mapping that groups words in the corpus based on unique bigrams from the beginning or ending of the words in the corpus;
- and
- replacing, at the computing device, the token with the best word.
A method, comprising:
- obtaining, at a computing device, optical character recognition (OCR) text extracted from a document image, the OCR text comprising a token;
- searching, at the computing device, a corpus for a best word to replace the token by performing operations comprising:
- determining a first set of one or more homoglyphs from a first token part of the token or a last token part of the token;
- creating a first list of one or more homoglyphs from the first set;
- determining a second set of one or more homoglyphs from a second token part of the token or a penultimate token part of the token;
- appending or prepending each of the homoglyphs of the second set to at least one of the homoglyphs of the first list to create an updated first list;
- and
- using a mapping that groups words in the corpus based on unique bigrams from the beginning or ending of the words in the corpus;
- and
- replacing, at the computing device, the token with the best word.
A non-transitory computer-readable storage medium storing instructions that, when executed by a processor, cause a system to perform operations comprising:
- obtaining text from a document associated with an application, the text comprising a token;
- obtaining a token bigram from the token by performing operations comprising:
- determining one or more homoglyph pairs of each of a first token part and a second token part of the token, or each of a last token part and a penultimate token part of the token;
- and
- determining the token bigram as one of the one or more homoglyph pairs;
- obtaining a plurality of possible words for the token by applying a bigram to corpus mapping filter to determine if the token bigram is the same as a bigram at the beginning of or the end of at least one word of a corpus;
- applying a probability filter to identify a word in the corpus as a most probable word of the possible words for the token;
- and
- replacing the token with the most probable word.
Dependencies are inferred from the claim text.