Form Value Extraction Using Hybrid Rule Configuration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing information processing systems face challenges in efficiently extracting item values from forms, as they require precise layout information and text recognition, which can be complex and user-dependent.
Innovation Solution
The system allows users to set rules for extracting item values from forms using a combination of predetermined rules, user-input rules based on element arrangement, and user-input coordinates, enabling flexible and accurate value extraction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If layout information and text recognition are used to extract item values from forms, then extraction accuracy is improved, but system complexity and user dependency increase
Solution Approach 1:
The patent divides the rule selection process into three distinct segments: predetermined rules for standard forms, user-input arrangement rules for customized layouts, and coordinate-based rules for precise positioning. This segmentation allows the system to handle different extraction scenarios with appropriate methods, reducing overall system complexity while maintaining high extraction accuracy.
Solution Approach 2:
The patent implements preliminary action by providing predetermined extraction rules that are pre-configured for common form types. Users can directly utilize these pre-established rules without performing complex layout analysis or text recognition, thereby simplifying the system while maintaining accuracy for standard forms.
2Measurement precision
If text recognition and position specification are required to extract item values, then extraction precision is improved, but ease of operation deteriorates
Solution Approach 1:
The patent enables self-service extraction by allowing users to define simple arrangement rules based on element positions and relationships without requiring complex text recognition or manual position specification. The system automatically applies these rules to extract item values, making the process user-friendly while maintaining precision.
Solution Approach 2:
The patent changes the parameter representation from complex text recognition results to simple arrangement parameters such as element positions, distances, and relative relationships. This parameter transformation maintains extraction precision while significantly improving ease of operation.
3Measurement precision
If structural dictionaries and layout information are used for aggregation, then extraction accuracy is improved, but adaptability to different form layouts deteriorates
Solution Approach 1:
The patent implements dynamics by making the extraction rules adaptable to different form layouts. Instead of using fixed structural dictionaries, the system allows users to define arrangement rules that can dynamically adjust to various form structures, and provides coordinate-based rules for precise positioning when needed, thereby maintaining both accuracy and adaptability.
Solution Approach 2:
The patent creates a universal extraction framework that can handle multiple form layouts through three types of rules. The predetermined rules cover standard forms, arrangement rules adapt to customized layouts, and coordinate rules provide precise extraction for any layout, making the system versatile while maintaining high extraction accuracy across different form types.
Data Source
AI summary
An information processing apparatus includes a processor configured to: set a rule for specifying a region in which a value of an item of an attribute assigned to a form shown by an image is shown in the image for each item, from a predetermined first rule, a second rule indicating an arrangement of the region for an element included in the form, which is input by a user, and a third rule indicating coordinates of the region in the image, which are input by the user; and extract the value of the item shown in the region specified by using the set rule.