Image Region Attribute Determination Using Hierarchical Statistical Accumulation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image processing technologies face challenges in accurately determining the attributes of target regions in images, leading to potential misclassification of image regions such as photos, text, and drawings.

Innovation Solution

An image processing device employing a multifunction peripheral with a CPU, storage, and scanning units, utilizing edge detection filters, region classification, and tree structure analysis to determine and modify attributes based on pixel density, brightness distribution, and color number, ensuring accurate identification of image regions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If image regions are divided into parent and child regions and attributes are determined separately for each region, then the flexibility to handle different region types is improved, but the accuracy of attribute determination deteriorates due to insufficient data for each individual region

Engineering Contradiction:
Improveflexibility to handle different region typesVSAvoidaccuracy of attribute determination
Core Design Contradiction:
Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent merges the attribute determination of parent regions and child regions by accumulating statistical information (pixel density, brightness distribution, color number) across multiple regions. The attribute of a parent region is determined based on aggregated data from its child regions, allowing accurate classification even when individual child regions have insufficient data.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent performs preliminary attribute determination on child regions before determining the attribute of parent regions. Statistical information is accumulated in advance from child regions to parent regions, enabling accurate attribute determination at higher levels without requiring separate analysis of each region type.

Inventive Principle:
Principle #10Preliminary action

2Measurement precision

If attribute determination is performed on each child region independently, then the detail level of region analysis is improved, but the computational complexity and time required increase

Engineering Contradiction:
Improvedetail level of region analysisVSAvoidcomputational time for attribute determination
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent performs attribute determination on child regions first and stores their statistical information (pixel density, brightness distribution, color number) for later use. This preliminary action eliminates the need to re-analyze child regions when determining parent region attributes, significantly reducing computational time.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent combines the results of child region analysis with parent region analysis by accumulating statistical data. Instead of performing separate independent analyses, the system merges child region statistics into parent region determinations, reducing redundant computations and overall processing time.

Inventive Principle:
Principle #5Merging (Combining)

3Measurement precision

If statistical information is accumulated from child regions to determine parent region attributes, then the accuracy of parent region classification is improved, but the complexity of the attribute determination process increases

Engineering Contradiction:
Improveaccuracy of parent region classificationVSAvoidcomplexity of attribute determination process
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent creates a universal attribute determination process that works for both child regions and parent regions using the same statistical metrics (pixel density, brightness distribution, color number). The accumulation mechanism serves multiple purposes: it determines child region attributes individually and simultaneously aggregates data for parent region determination, simplifying the overall process architecture.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP2551795B1Image processing device determining attributes of regions
Publication Date: 2019.06.26 BROTHER KOGYO KK
  • EP2551795B1 patent drawingFigure 1
  • EP2551795B1 patent drawingFigure 2(a)
  • EP2551795B1 patent drawingFigure 2(b)

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: preparing image data representing an image; identifying a first region in the image and a second region disposed inside of the first region; determining an attribute of the first region to be one of a plurality of attributes; and determining, when the attribute of the first region is determined to be the first type attribute, an attribute of the second region by using the attribute of the first region. The plurality of attributes includes a first type attribute indicating one of photo and drawing.