Back to results
US 11,537,605 B2Granted
Method, apparatus, and system for auto-registration of nested tables from unstructured cell association for table-based documentation
AssigneeKonica Minolta Business Solutions U.S.A., Inc.
InventorJunchao Wei
Life of this patent
Estimating the term…
Abstract
In some forms containing keywords and content, there may be nested levels of keywords, also referred to as a hierarchy. Content in the forms may be associated with one or more keywords in one or more of the nested levels, or in the hierarchy. Identifying keywords in adjacent cells in a table (with a nested keyword being either to the right of or below another keyword) enables distinguishing between keywords and content in filled forms, and enables correct association of content with respective keywords.
Claims20
A form auto-registration method comprising:
- responsive to an input of a form, extracting one or more features from the form;
- using a deep learning model, identifying lines from other features in the form, and extracting the lines;
- using the deep learning model, identifying and extracting connected regions using the extracted lines;
- using the deep learning model, identifying a table from the connected regions;
- extracting cells from the identified table;
- grouping the cells of the identified table using locations and sizes of adjacent cells in the identified table;
- forming rows and columns in the identified table;
- forming a hierarchy for the cells in the identified table using the grouping and one of the formed rows and the formed columns;
- identifying adjacent keywords in the identified table using the formed hierarchy;
- identifying content in the identified table using the formed hierarchy;
- and
- associating respective identified adjacent keywords with respective identified content.
9 more elements
A form auto-registration system comprising:
- one or more processors;
- one or more non-transitory memory devices;
- and
- a deep learning system which implements a deep learning model;
- the auto-registration system storing one or more programs in the one or more non-transitory memory devices, the one or more programs containing instructions which, when executed, perform the following:
- responsive to an input of a form, extracting one or more features from the form;
- using a deep learning model, identifying lines from other features in the form, and extracting the lines;
- using the deep learning model, identifying and extracting connected regions using the extracted lines;
- using the deep learning model, identifying a table from the connected regions;
- extracting cells from the identified table;
- grouping the cells of the identified table using locations and sizes of adjacent cells in the identified table;
- forming rows and columns in the identified table;
- forming a hierarchy for the cells in the identified table using the grouping and one of the formed rows and the formed columns;
- identifying adjacent keywords in the identified table using the formed hierarchy;
- identifying content in the identified table using the formed hierarchy;
- and
- associating respective identified adjacent keywords with respective content.
14 more elements
Dependencies are inferred from the claim text.