Image Processing Apparatus Using Block Region Frequency Analysis

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing image processing methods struggle with accurate recognition of object types in images, leading to incorrect classifications and searches, as they rely solely on physical characteristics, which can confuse similar objects like skin and sand, resulting in low accuracy and unnatural image textures.

Innovation Solution

An image processing method that divides images into object regions and further into block regions, recognizes the types of block regions, totals their occurrence frequencies to determine object types, and sets image processing conditions based on type reliability values, using block characteristic quantities like color, lightness, and structural components to improve accuracy and prevent image degradation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If image classification and search are carried out based on physical characteristics alone, then processing speed is improved, but accuracy deteriorates due to misclassification of similar objects

Engineering Contradiction:

Improveprocessing speedVSAvoidclassification accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The image is divided into multiple block regions, and each block region is independently analyzed for type recognition. This segmentation allows the system to examine local characteristics while maintaining overall processing efficiency, resolving the contradiction between speed and accuracy by processing discrete segments rather than the entire image at once.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Instead of analyzing the entire image uniformly, the patent applies type recognition to selected block regions and uses occurrence frequency thresholds. This partial action approach focuses computational resources on critical areas, improving accuracy without sacrificing overall processing speed.

Inventive Principle:

Principle #16Partial or excessive action

2Object-affected harmful factors

If noise reduction is carried out on areas recognized by color alone, then noise removal effectiveness is improved, but image quality deteriorates due to texture distortion

Engineering Contradiction:

Improvenoise removal effectivenessVSAvoidimage texture quality

Core Design Contradiction:

Object-affected harmful factorsVSManufacturing precision

Solution Approach 1:

The patent performs type recognition on block regions before applying noise reduction processing. By preliminarily identifying the type of each block region (e.g., skin, sand, water), the system can then apply appropriate noise reduction strategies that preserve texture characteristics, preventing the distortion that occurs when color-based recognition alone is used.

Inventive Principle:

Principle #10Preliminary action

3Ease of operation

If automatic type recognition is implemented, then ease of operation is improved, but device complexity increases due to additional processing steps

Engineering Contradiction:

Improveautomation levelVSAvoidprocessing system complexity

Core Design Contradiction:

Ease of operationVSDevice complexity

Solution Approach 1:

The patent creates a universal type recognition system that can identify multiple object types (skin, sand, water, sky, building, tree, mountain, flower, animal, vehicle, food, other) using a unified approach based on block region analysis and occurrence frequency. This multi-functional system handles diverse recognition tasks through a single mechanism, improving ease of operation without proportionally increasing complexity.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The system automatically performs type recognition and determines image processing conditions without user intervention. The automatic type recognition unit independently analyzes block regions, calculates occurrence frequencies, and sets processing conditions, enabling the system to serve itself and eliminating the need for manual object region extraction and type input.

Inventive Principle:

Principle #25Self-service

4Measurement precision

If block region analysis is performed for type recognition, then measurement precision is improved, but use of energy increases due to additional processing

Engineering Contradiction:

Improvetype recognition accuracyVSAvoidprocessing energy consumption

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent analyzes only selected block regions rather than every pixel in the image, and uses occurrence frequency thresholds to determine types. This partial analysis approach achieves high recognition accuracy while significantly reducing the computational energy required compared to exhaustive image analysis.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS7620246B2Method and apparatus for image processing

Publication Date: 2009.11.17 FUJIFILM CORP

AI summary

The type of an object included in an image is automatically recognized. An image is divided into object regions and into block regions each having a predetermined number of pixels and smaller than any one of the object regions. The types of the respective block regions are recognized and totaled up for each of the object regions. The type of each of the object regions is then recognized by using a result of totaling.