Document Image Extraction Using Frequency-Based Area Analysis

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing image processing systems face difficulties in accurately extracting document images from images with backgrounds that are not clearly distinguishable from the document colors, particularly when using non-black backing sheets, leading to incomplete or incorrect detection of document areas.

Innovation Solution

An image processing apparatus that includes a first extraction unit for low-frequency characteristics, a second extraction unit for high-frequency characteristics, and a document area estimation module that combines these extractions based on whether the background is white, allowing for precise extraction of document images regardless of the background color.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a black backing sheet is used to improve document extraction accuracy, then measurement precision improves, but device complexity and ease of operation worsen due to mandatory black background requirement

Engineering Contradiction:

Improvedocument extraction accuracyVSAvoidoperation flexibility

Core Design Contradiction:

Measurement precisionVSEase of operation

Solution Approach 1:

The patent changes the parameter of background color from fixed (black only) to variable (any color), allowing the system to adapt to different backing sheet colors while maintaining accurate document extraction through color-based differentiation algorithms

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

Instead of requiring the background to be black to enable extraction, the patent inverts the approach by enabling extraction of any document color against any background color through advanced image processing that identifies document regions based on color patterns rather than assuming a fixed black background

Inventive Principle:

Principle #13The other way round (Inversion)

2Device complexity

If simple extraction methods are used to maintain device simplicity, then device complexity is reduced, but measurement precision deteriorates when backgrounds are not clearly distinguishable

Engineering Contradiction:

Improveprocessing system simplicityVSAvoiddocument area detection accuracy

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

The patent introduces dynamic adjustment of extraction parameters based on the detected background color, allowing the system to automatically adapt its processing strategy to match the specific characteristics of each image, thereby maintaining high accuracy without requiring complex fixed-rule systems

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system employs feedback mechanisms where the detected background color informs subsequent extraction parameter selection, creating a closed-loop process that continuously optimizes extraction accuracy based on actual image conditions while maintaining relatively simple overall system architecture

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS10318801B2Image processing apparatus and non-transitory computer readable medium

Publication Date: 2019.06.11 FUJIFILM BUSINESS INNOVATION CORP

AI summary

An image processing apparatus includes a first extraction unit, a second extraction unit, and a third extraction unit. The first extraction unit extracts a first area in an image that includes plural document images, the first area having low-frequency characteristics. The second extraction unit extracts a second area in the image, the second area having high-frequency characteristics. The third extraction unit extracts areas of the document images by combining the first area and the second area in accordance with whether a background of the image is white.