Character Rectangle Extraction for Overlapping Text Recognition

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing character recognition techniques face a decrease in accuracy when characters close to each other overlap, leading to the removal of one character or extraction of unnecessary character region rectangles.

Innovation Solution

An image processing apparatus that includes an acquisition processing unit, detection processing unit, correction processing unit, and extraction processing unit to improve character recognition accuracy by correcting the reliability of overlapping rectangle candidates based on the type of characters using specific overlap calculation methods.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If character region rectangles are extracted using conventional overlap removal methods, then processing complexity is reduced, but character recognition accuracy decreases

Engineering Contradiction:

Improveprocessing complexityVSAvoidcharacter recognition accuracy

Core Design Contradiction:

Device complexityVSReliability

Solution Approach 1:

The patent changes the parameter being optimized from simple rectangle extraction to reliability-weighted rectangle extraction. By introducing a reliability calculation that considers overlap relationships between character regions, the system adjusts which rectangles are extracted based on their reliability scores rather than applying fixed removal rules. This allows the system to maintain higher recognition accuracy while managing processing complexity through intelligent parameter adjustment.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent replaces the mechanical overlap removal approach with a reliability-based selection mechanism. Instead of mechanically removing overlapping rectangles based on geometric rules, the system substitutes this with a computational reliability assessment that evaluates each rectangle candidate and selects the most reliable ones, thereby improving accuracy without excessive complexity.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Productivity

If overlapping rectangle candidates are removed to simplify processing, then processing speed increases, but character recognition accuracy decreases

Engineering Contradiction:

Improveprocessing speedVSAvoidcharacter recognition accuracy

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The system changes the decision parameter from binary removal/extraction to reliability-based weighted selection. By calculating reliability scores for overlapping rectangle candidates and using these scores to determine extraction decisions, the system achieves better accuracy while maintaining reasonable processing speed through efficient reliability computation rather than exhaustive analysis.

Inventive Principle:

Principle #35Parameter changes

3Reliability

If all detected rectangle candidates are extracted, then character recognition coverage improves, but unnecessary extractions increase

Engineering Contradiction:

Improvecharacter recognition coverageVSAvoidunnecessary extractions

Core Design Contradiction:

ReliabilityVSLoss of information

Solution Approach 1:

The patent replaces the mechanical approach of extracting all detected rectangles with a reliability-based filtering mechanism. By substituting blind extraction with reliability assessment, the system identifies and extracts only the most reliable character region rectangles, thereby improving recognition coverage while eliminating unnecessary extractions that would degrade information quality.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The reliability calculation serves as an intermediary between detection and extraction. Rather than directly extracting all detected rectangles or removing overlapping ones, the system introduces reliability scores as an intermediate evaluation layer that mediates the extraction decision, allowing the system to select appropriate rectangles while filtering out unnecessary ones.

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS20260065624A1Image processing apparatus, image processing system, output apparatus, image processing method, and recording medium in which image processing program is recorded

Publication Date: 2026.03.05 SHARP KK

AI summary

The display device includes an acquisition processing unit, a rectangle candidate detection processing unit, and a one-character extraction processing unit. The acquisition processing unit acquires image data including characters. The rectangle candidate detection processing unit detects a rectangle candidate of a character from the image data, and calculates a reliability of the detected rectangle candidate. The one-character extraction processing unit corrects, in a case where a first rectangle candidate and a second rectangle candidate detected by the detection processing unit at least partially overlap each other, a reliability of a rectangle candidate to be corrected out of the first rectangle candidate and the second rectangle candidate by using a calculation method for a degree of overlap selected according to a relationship between a type of a first detection target corresponding to the first rectangle candidate and a type of a second detection target corresponding to the second rectangle candidate. The one-character extraction processing unit extracts a one-character rectangle based on the corrected reliability.