Image Processing Apparatus Intermittent Line Classification

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing image processing systems fail to accurately classify intermittent lines, often misclassifying them as characters, leading to ineffective character recognition results.

Innovation Solution

An image processing apparatus utilizing a generative adversarial network (GAN) modifies intermittent lines into solid lines before classification, ensuring they are correctly identified as marks rather than characters, thereby preventing misclassification.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If machine learning is applied to classify image elements into characters and marks, then classification capability is improved, but intermittent lines are misclassified as characters reducing reliability

Engineering Contradiction:

Improveclassification capabilityVSAvoidclassification accuracy

Core Design Contradiction:

Extent of automationVSReliability

Solution Approach 1:

The patent applies preliminary action by modifying the intermittent line image data before classification occurs. The GAN-based modifying unit transforms intermittent lines into solid lines in advance, so that when the classifying unit processes the data, the lines are already in a recognizable solid format that correctly identifies them as marks rather than characters.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent introduces an intermediary element - the GAN-based modifying unit - that acts as a bridge between the raw image data and the classification process. This intermediary transforms the problematic intermittent line format into a standardized solid line format, enabling the classifier to function reliably without directly handling the ambiguous intermittent line cases.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Productivity

If intermittent lines are processed as-is, then processing speed is maintained, but misclassification occurs reducing productivity

Engineering Contradiction:

Improvecharacter recognition efficiencyVSAvoidclassification accuracy

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The system performs preliminary modification of intermittent lines to solid lines before the classification stage, ensuring that the main character recognition process can proceed efficiently without needing to handle ambiguous intermittent line cases separately, thus maintaining overall productivity while improving accuracy.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent replaces traditional rule-based or manual classification methods with a GAN-based deep learning system that automatically learns to distinguish and transform intermittent lines, substituting complex mechanical classification rules with an intelligent system that achieves both speed and accuracy.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS11611678B2Image processing apparatus and non-transitory computer readable medium

Publication Date: 2023.03.21 FUJIFILM BUSINESS INNOVATION CORP

AI summary

An image processing apparatus includes an acquisition unit that acquires an image; and a modifying unit that modifies the image acquired by the acquisition unit by turning an intermittent line different from a line that constitutes a character into a mark by using machine learning in a stage before the image is classified into the character and a mark by a classifying unit.