Dynamic File Naming via Handwritten Mark Detection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing image processing systems are not suitable for setting file names based on user-selected character strings from business forms with multiple options, as they rely on pre-learned text block positions, leading to constant use of the same character string for all forms of the same type.

Innovation Solution

An image processing apparatus that learns the position of user-selected character regions, including both printed and handwritten areas, to determine the correct file name for similar documents by identifying nearby handwritten and printed regions, allowing dynamic selection of character strings based on user input.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If the text block located at the same position is used for the file_name of each business form of the same type, then the file_name setting process is simplified and automated, but the system cannot correctly identify user-selected character strings when options change on business forms

Engineering Contradiction:

Improveautomation of file_name settingVSAvoidadaptability to user-selected options

Core Design Contradiction:

Extent of automationVSAdaptability or versatility

Solution Approach 1:

The system dynamically adjusts the text block selection based on handwritten markings. Instead of using a fixed position, the system identifies the position of handwritten character strings (circles, checkmarks) and selects the text block corresponding to the marked option, enabling the system to adapt automatically to user selections while maintaining automation.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system uses the handwritten markings on the business form as feedback to determine which text block should be used for the file_name. The presence and position of handwritten symbols provide information that guides the selection process, allowing the system to correctly identify user intent without manual intervention.

Inventive Principle:

Principle #23Feedback

2Ease of manufacture

If only the position of a character string corresponding to an option designated as the file_name is learned, then the learning process is simple, but the system fails to identify other options when users select different options on similar business forms

Engineering Contradiction:

Improvesimplicity of learning processVSAvoidability to handle multiple options

Core Design Contradiction:

Ease of manufactureVSAdaptability or versatility

Solution Approach 1:

The system segments the business form into multiple text blocks, each corresponding to a different option. Instead of learning a single text block position, the system identifies and stores information about multiple text blocks associated with handwritten markings, allowing it to select the appropriate one based on user selection while maintaining a simple learning process.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The learning process is designed to handle multiple functions: it can learn text block positions for any option on the business form, not just a predetermined one. The system stores information about multiple text blocks and their associations with handwritten markings, making it universally applicable to any option selection scenario.

Inventive Principle:

Principle #6Universality (Multi-functionality)

3Productivity

If the system constantly uses the character string located at the learned position as the file_name, then processing is efficient and quick, but the file_name does not reflect the actual user selection on the business form

Engineering Contradiction:

Improveprocessing speedVSAvoidaccuracy of file_name extraction

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The system performs preliminary identification of handwritten markings and their positions on the business form before extracting the file_name. By pre-identifying the location of user selections (circles, checkmarks), the system can quickly extract the correct text block without trial and error, maintaining high processing speed while ensuring accuracy.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system replaces the mechanical approach of using fixed-position text extraction with an intelligent system that recognizes handwritten markings and dynamically determines text block positions. This substitution of recognition-based methods for position-based methods maintains efficiency while dramatically improving accuracy in identifying user selections.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS11620840B2Image processing apparatus for extracting a desired character string from a scanned image

Publication Date: 2023.04.04 CANON KK

AI summary

In the case of learning a printed character region selected by a user, it is determined whether a handwritten character region is present near the printed character region. If it is determined that the handwritten character region is present near the printed character region, information about the handwritten character region present near the printed character region and information about another printed character region present near the printed character region are learned in association with information about the selected printed character region. This makes it possible to appropriately determine a circled character string and a character string selected with a check mark from among a plurality of options during a scanned image analysis.