Back to results
US 10,853,640 B2Granted
Method and system for extracting information from hand-marked industrial inspection sheets
AssigneeTata Consultancy Services Limited
InventorsGaurav Gupta, Swati, Monika Sharma, Lovekesh Vig
Life of this patent
Estimating the term…
Abstract
This disclosure relates generally to document processing, and more particularly to extracting information from hand-marked industrial inspection sheets. In an embodiment, the system performs localization of text as well as arrows in the inspection sheet, and identifies text that matches each arrow. Further by identifying machine zone each arrow is pointing to, the system assigns corresponding text to the appropriate machine zone; thus facilitating digitization of the inspection sheets.
Claims17
A processor-implemented method, comprising:
- collecting an inspection sheet as input, via one or more hardware processors, by a document digitization system;
- performing, via the one or more hardware processors, a component localization of one or more line diagrams in the inspection sheet, by the document digitization system, wherein the component localization involves identification and classification of different machine zones in the one or more line diagrams, wherein the component localization comprises:
- performing a pixel level comparison of the inspection sheet with a plurality of template sheets stored in a database associated with the document digitization system, wherein said plurality of template sheets have one or more line diagrams with zones represented as contours;
- identifying at least one template sheet as matching with the inspection sheet;
- and
- identifying different machine zones in the inspection sheet, based on the contours defined in the template sheet that is identified as the template sheet matching with the inspection sheet;
- performing, via the one or more hardware processors, an arrow head-tail localization of the one or more line diagrams in the inspection sheet, by the document digitization system, wherein the arrow head-tail localization involves identifying arrow head and tail points in the inspection sheet, for all arrows present in the inspection sheet;
- performing, via the one or more hardware processors, text detection in the inspection sheet, by the document digitization system, wherein the text detection involves detection of all texts in the inspection sheet, and position of identified text with respect to one or more clusters in the inspection sheet;
- performing, via the one or more hardware processors, a text to arrow mapping, by the document digitization system, wherein the text to arrow mapping is based on the identified head and tail points of arrows and position of cluster in which the text is present;
- performing, via the one or more hardware processors, a zone to arrow mapping, by the document digitization system, wherein the zone to arrow mapping is based on the identified head and tail points of arrows and machine zones identified during the component localization;
- and
- identifying text corresponding to different machine zones, based on result of the text to arrow mapping and zone to arrow mapping, via the one or more hardware processors, by the document digitization system.
9 more elements
A document digitization system, comprising:
- a processor;
- and
- a memory module comprising a plurality of instructions, said plurality of instructions configured to cause the processor to:
- collect an inspection sheet as input, via one or more hardware processors, by an Input/Output (I/O) module of the document digitization system;
- perform a component localization of one or more line diagrams in the inspection sheet, by a component localization module of the document digitization system, wherein the component localization involves identification and classification of different machine zones in the one or more line diagrams, wherein the component localization comprises:
- performing a pixel level comparison of comparing the inspection sheet with a plurality of template sheets stored in a database associated with the document digitization system, wherein said plurality of template sheets have one or more line diagrams with zones represented as contours:
- identifying at least one template sheet as matching with the inspection sheet;
- and
- identifying different machine zones in the inspection sheet, based on the contours defined in the template sheet that is identified as the template sheet matching with the inspection sheet;
- perform an arrow head-tail localization of the one or more line diagrams in the inspection sheet, by an arrow localization module of the document digitization system, wherein the arrow head-tail localization involves identifying arrow head and tail points in the inspection sheet, for all arrows present in the inspection sheet;
- perform text detection in the inspection sheet, by a text detection module of the document digitization system, wherein the text detection involves detection of all texts in the inspection sheet, and position of identified text with respect to one or more clusters in the inspection sheet;
- perform a text to arrow mapping, by a mapping module of the document digitization system, wherein the text to arrow mapping is based on the identified head and tail points of arrows and position of cluster in which the text is present;
- perform a zone to arrow mapping, by the mapping module, wherein the zone to arrow mapping is based on the identified head and tail points of arrows and machine zones identified during the component localization;
- and
- identify text corresponding to different machine zones, based on result of the text to arrow mapping and zone to arrow mapping, by the mapping module.
12 more elements
One or more non-transitory machine readable information storage mediums comprising one or more instructions which when executed by one or more hardware processors causes:
- collecting an inspection sheet as input;
- performing a component localization of one or more line diagrams in the inspection sheet, wherein the component localization involves identification and classification of different machine zones in the one or more line diagrams, wherein the component localization comprises:
- performing a pixel level comparison of the inspection sheet with a plurality of template sheets stored in a database associated with the document digitization system, wherein said plurality of template sheets have one or more line diagrams with zones represented as contours;
- identifying at least one template sheet as matching with the inspection sheet;
- and
- identifying different machine zones in the inspection sheet, based on the contours defined in the template sheet that is identified as the template sheet matching with the inspection sheet;
- performing an arrow head-tail localization of the one or more line diagrams in the inspection sheet, wherein the arrow head-tail localization involves identifying arrow head and tail points in the inspection sheet, for all arrows present in the inspection sheet;
- performing text detection in the inspection sheet, wherein the text detection involves detection of all texts in the inspection sheet, and position of identified text with respect to one or more clusters in the inspection sheet;
- performing a text to arrow mapping, wherein the text to arrow mapping is based on the identified head and tail points of arrows and position of cluster in which the text is present;
- performing a zone to arrow mapping, wherein the zone to arrow mapping is based on the identified head and tail points of arrows and machine zones identified during the component localization;
- and
- identifying text corresponding to different machine zones, based on result of the text to arrow mapping and zone to arrow mapping.
9 more elements
Dependencies are inferred from the claim text.