Image Processing Edge Detection for Region Separation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image processing methods struggle to accurately separate regions with similar color differences, such as text and background, when converting images to binary format, leading to poor precision in identifying different types of regions within an image.

Innovation Solution

An image processing device that generates edge image data using the Sobel operator to calculate edge strength, determines uniform regions by averaging pixel values across target blocks, and consolidates pixels with similar flags to identify and label regions, improving separation accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If image is converted to binary format for region separation, then processing speed is improved, but measurement precision deteriorates when regions have similar color differences

Engineering Contradiction:
Improveprocessing speedVSAvoidregion separation precision
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent applies preliminary action by performing grayscale conversion and edge detection before binary conversion. The edge detection step identifies boundary characteristics in grayscale images, preserving structural information that would be lost in direct binary conversion. This preliminary processing enables accurate region separation even when color differences are minimal, as the edge information provides additional discrimination cues before the thresholding operation.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent employs parameter changes by transitioning from simple binary thresholding to a multi-parameter approach that considers edge strength, gradient magnitude, and local variance in addition to intensity values. By changing the parameters used for region separation from单一的intensity threshold to multiple characteristics including edge properties, the system maintains high processing speed while significantly improving separation precision for regions with similar colors.

Inventive Principle:
Principle #35Parameter changes

2Device complexity

If binary conversion is used for region identification, then device complexity is reduced, but manufacturing precision deteriorates in separating regions with similar display content

Engineering Contradiction:
Improveprocessing complexityVSAvoidregion identification precision
Core Design Contradiction:
Device complexityVSManufacturing precision

Solution Approach 1:

The patent applies segmentation by dividing the image processing into distinct stages: grayscale conversion, edge detection, and then binary conversion for final region identification. This segmentation allows each stage to specialize in extracting specific features (intensity, then edge characteristics, then binary regions), maintaining low overall device complexity while achieving high precision through the cumulative effect of staged processing.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces an intermediary edge detection stage between grayscale conversion and binary conversion. This intermediary process acts as a mediator that enhances boundary information without requiring complex hardware. The edge detection algorithm serves as a computational intermediary that bridges the simple grayscale-to-binary conversion while adding the necessary precision for separating regions with similar display content.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP2551823B1Image processing device identifying region in image as one of uniform region and nonuniform region
Publication Date: 2019.09.25 BROTHER KOGYO KK
  • EP2551823B1 patent drawingFigure 1
  • EP2551823B1 patent drawingFigure 2~4
  • EP2551823B1 patent drawingFigure 5

AI summary

An image processing device includes a processor, and a memory storing computer-readable instructions therein. The computer-readable instructions, when executed by the processor, causes the image processing device to perform: generating edge image data by using the original image data; calculating characteristic values for a plurality of determination regions; and identifying a determination region as a nonuniform region when the characteristic value of the determination region satisfies a prescribed criterion, and the determination region as a uniform region when the characteristic value of the determination region does not satisfy the prescribed criterion. Each of the plurality of determination regions corresponds to one of the characteristic values, represents a part of the edge image, and includes a plurality of pixels, the plurality of determination regions being different from one another, each of the characteristic values characterizing the edge strength of the corresponding determination region.