Image Processing Apparatus for Separating Overlapping Documents

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional image processing apparatuses struggle to separate overlapping original documents during the multi-cropping process, often cutting out multiple documents as a single entity.

Innovation Solution

An image processing apparatus that extracts contour lines from scanned images, calculates vertices to identify document areas, and forms rectangles to isolate individual documents, even when they overlap, allowing for accurate separation and generation of separate files for each document.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If conventional multi-cropping processing is used to scan multiple original documents at once, then processing efficiency is improved, but overlapping documents cannot be separated and are cut out as a single document

Engineering Contradiction:
Improveprocessing efficiencyVSAvoiddocument separation accuracy
Core Design Contradiction:
ProductivityVSManufacturing precision

Solution Approach 1:

The patent segments the scanned image into multiple document regions by detecting contour lines and identifying vertices. The processor divides the image data into separate document areas based on geometric features (vertices forming rectangles), enabling individual extraction of each document even when they overlap in the scan.

Inventive Principle:
Principle #1Segmentation

2Manufacturing precision

If the image processing apparatus attempts to separate overlapping documents, then document separation accuracy is improved, but the device complexity increases due to vertex calculation and rectangle formation algorithms

Engineering Contradiction:
Improvedocument separation accuracyVSAvoidprocessing algorithm complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent replaces complex image analysis methods with geometric calculations based on vertex detection. Instead of using sophisticated machine learning or pattern recognition algorithms, the system calculates curvature at contour points to identify vertices, then forms rectangles through mathematical operations on these vertices to define document boundaries.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Measurement precision

If vertex extraction is performed by calculating curvature at each contour point to identify document corners, then document area identification accuracy is improved, but processing time increases

Engineering Contradiction:
Improvevertex detection accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent performs preliminary contour line extraction from the scanned image data before vertex detection. By first identifying the contour lines that define document boundaries, the system narrows down the search space for vertex detection to only the relevant edge points, reducing the overall processing time while maintaining accurate vertex identification through curvature calculation.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3438931B1Image processing apparatus
Publication Date: 2022.02.23 KK TOSHIBA
  • EP3438931B1 patent drawingFigure 1~2
  • EP3438931B1 patent drawingFigure 3~4
  • EP3438931B1 patent drawingFigure 5~6

AI summary

An image processing apparatus includes an interface through which scanned image data obtained by scanning a designated region, is acquired, and a processor. The processor extracts a contour line representing a document area from the scanned image data, extracts vertices of the contour line, determines whether a plurality of original documents is present in the document area based on the vertices, extracts document image data of each original document from the scanned image data, and generates one file including the extracted document image data for each original document.