Learning image templates for content anchoring and data extraction
AssigneePalo Alto Research Center Incorporated
InventorPrateek Sarkar
Life of this patent
Estimating the term…
Abstract
Methods, and corresponding systems, of generating one or more image anchor templates for extracting data from a data field of a first class of documents are provided. The methods include generating one or more candidate image anchor templates from at least one of one or more exemplars of the first class; determining a quality score for each of the one or more candidate image anchor templates using a computer processor and known locations of the data field within the one or more exemplars of the first class; ranking the one or more candidate image anchor templates according to quality score; and selecting one or more of the most highly ranked image anchor templates.
Claims20
A method of generating one or more image anchor templates for extracting data from a data field of a first class of documents, said method comprising:
- a) generating one or more candidate image anchor templates using one or more seed image anchor templates and/or at least one of one or more exemplars of the first class;
- b) determining a quality score for each of the one or more candidate image anchor templates using a computer processor and known locations of the data field within the one or more exemplars of the first class;
- c) ranking the one or more candidate image anchor templates according to quality score;
- and
- d) selecting one or more of the most highly ranked image anchor templates.
A system of generating one or more image anchor templates for extracting data from a data field of a first class of documents, said system comprising:
- a generator module that generates one or more candidate image anchor templates from at least one of one or more exemplars of the first class;
- a scoring module that determines a quality score for each of the one or more candidate image anchor templates using a computer processor and known locations of the data field within the one or more exemplars of the first class;
- a ranking module that ranks the one or more candidate image anchor templates according to quality score;
- and
- a selection module that selects a one or more of the most highly ranked image anchor templates.
A method of extracting data from a data field of a document belonging to a first class of documents, said method comprising:
- a) generating one or more candidate image anchor templates from at least one of one or more exemplars of the first class;
- b) determining a quality score for each of the one or more candidate image anchor templates using a computer processor and known locations of the data field within the one or more exemplars of the first class;
- c) ranking the one or more candidate image anchor templates according to quality score;
- d) selecting one or more of the most highly ranked image anchor templates;
- e) predicting a location of the data field for each of the selected image anchor templates using a match location of the each of the selected image anchor templates to the document;
- and
- f) extracting data from the data field if the predicted locations of the data field for the selected image anchor templates have low variance.
Dependencies are inferred from the claim text.