Receipt Text Extraction via Relative Position and Attribute Rules

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The existing technologies for processing data from POS terminals are cumbersome, requiring users to perform complex operations for extracting text information from printed receipts, which is time-consuming and inefficient.

Innovation Solution

An information processing device and method that simplifies the extraction of text data by allowing users to generate extraction conditions through a user-friendly interface, enabling the selection of text related to extraction items and setting item values from a menu, thereby streamlining the text extraction process.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If text extraction is performed by existing technologies, then text data can be extracted from printed receipts, but the operation becomes cumbersome and time-consuming

Engineering Contradiction:
Improvetext extraction efficiencyVSAvoiduser operation complexity
Core Design Contradiction:
ProductivityVSEase of operation

Solution Approach 1:

The system performs automatic text extraction without requiring manual user intervention. The information processing device automatically identifies and extracts text data from printed receipts based on predetermined extraction conditions, eliminating the need for users to manually perform complex extraction operations.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The system pre-configures extraction conditions and text data structures before the actual extraction process. By preparing extraction rules, field mappings, and data formats in advance, the system enables rapid text extraction when receipts are processed, avoiding time-consuming manual configuration during operation.

Inventive Principle:
Principle #10Preliminary action

2Measurement precision

If complex extraction operations are required, then precise text extraction can be achieved, but the operational burden on users increases

Engineering Contradiction:
Improvetext extraction accuracyVSAvoidextraction process complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system introduces an intermediary processing layer that automatically handles the complex extraction operations. This intermediary layer includes preprocessing modules that prepare the receipt data, extraction modules that apply complex extraction rules, and postprocessing modules that refine the extracted text, thereby shielding users from complexity while maintaining high extraction accuracy.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The text extraction process is divided into multiple independent modules: preprocessing module for preparing receipt images, extraction module for applying extraction rules, and postprocessing module for refining results. Each module handles specific tasks with high precision, and their combination achieves accurate text extraction without requiring users to manage the overall complexity.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3489816B1Information processing device, control method of information processing device, and storage medium
Publication Date: 2024.02.28 SEIKO EPSON CORP
  • EP3489816B1 patent drawingFigure 1
  • EP3489816B1 patent drawingFigure 2
  • EP3489816B1 patent drawingFigure 3

AI summary

A control server which includes a control server controller which, on a screen for displaying print text data according to a layout according to which receipt information is printed on a roll paper, displays a plurality of setting items corresponding to an extraction item as a setting menu and receives selection of the setting item in a state where selection of extraction item related text is received and the extraction item related text is selected, sets the extraction item related text as item value text in a case where an item value setting item is selected, generates an extraction condition for extracting the item value text from print text data based on at least one of a relative position of the item value text in the print text data and an attribute of the item value text, and extracts text matching the generated extraction condition as the item value text.