System and method for extracting information from documents
AssigneeFirst American Data Tree LLC
InventorsChristopher Lawrence Rubio, Vladimir Sevastyanov
Life of this patent
Estimating the term…
Abstract
Various embodiments of the invention provide systems and methods for extracting information from digital documents, including physical documents that have been converted to digital documents. For example, some embodiments are configured to extract information from a field in a digital document by identifying a block of tokens before (i.e., a prior block) and a block of tokens after (i.e., a post block) the field from which the information is to be extracted, where both the prior block and post block are known to be associated with the field type of the field (e.g., name, address, phone number, etc.).
Claims19
A computer program product for extracting field data from a digital real estate document, the product comprising a non-transitory computer-readable medium embedded therein, the program instructions configured to cause a computer system to perform the operations of:
- using an extraction field template for a document classification to identify in the digital real estate document a set of candidate prior blocks for a specific field type and a set of candidate post blocks for the specific field type;
- calculating a confidence score for a candidate prior block in the set of candidate prior blocks and a confidence score for a candidate post block in the set of candidate post blocks;
- selecting a best prior block from the set of candidate prior blocks based on the best prior block's confidence score and a best post block from the set of candidate post blocks based on the best post block's confidence score;
- and
- extracting a token block located between the best candidate prior block and the best candidate post block, wherein the token block is the field data and is associated with the specific field type.
Dependencies are inferred from the claim text.