Rectangle Region Detection in Image Processing Apparatus

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image processing systems face challenges in accurately detecting rectangle regions from images, especially when documents are scanned without a backing, leading to incorrect detection due to illumination and other factors.

Innovation Solution

An image processing apparatus that extracts edge pixels, adds end pixels based on pixel values, and detects rectangle regions using straight lines, incorporating edge pixel deletion and addition processes to improve accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of time

If a document is read without a backing (cover open), then operation time is shortened, but illumination and other factors are captured in regions not containing document, hindering accurate detection of document region

Engineering Contradiction:
Improveoperation timeVSAvoiddocument region detection accuracy
Core Design Contradiction:
Loss of timeVSMeasurement precision

Solution Approach 1:

The patent segments the image processing into multiple stages: edge pixel extraction, straight line extraction from edge pixels, rectangle region candidate formation, and selection based on evaluation criteria. This segmentation allows the system to handle images without backings by systematically identifying document regions through hierarchical processing steps.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent changes the detection parameters by evaluating rectangle region candidates based on the number of edge pixels within a prescribed distance from each side and the likelihood of corners. This parameter-based evaluation enables accurate document region detection even when illumination varies or no backing is present, as it focuses on structural features rather than uniform background assumptions.

Inventive Principle:
Principle #35Parameter changes

2Device complexity

If edge pixels are extracted from the entire image including regions without document, then processing is simplified, but incorrect edge pixels are included leading to inaccurate rectangle region detection

Engineering Contradiction:
Improveprocessing complexityVSAvoidrectangle region detection accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent performs preliminary edge pixel extraction from the entire image, then subsequently filters and evaluates these edge pixels to form straight lines and rectangle candidates. This preliminary action approach maintains processing simplicity while enabling accurate filtering in later stages through systematic evaluation criteria for selecting valid rectangle regions.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent incorporates feedback mechanisms by evaluating rectangle region candidates based on multiple criteria: the number of edge pixels within prescribed distances from each side, and the likelihood of corner positions. This feedback-based selection process refines the initial edge pixel extraction results, eliminating incorrect detections while preserving valid document regions.

Inventive Principle:
Principle #23Feedback

Data Source

PatentUS10230867B2Image processing apparatus, rectangle region detection method, and computer-readable, non-transitory medium
Publication Date: 2019.03.12 PFU LTD
  • US10230867B2 patent drawing
  • US10230867B2 patent drawing
  • US10230867B2 patent drawing

AI summary

An image processing apparatus includes an edge pixel extractor for extracting an edge pixel from an input image, an end pixel extractor for extracting an end pixel in the input image, an edge pixel addition module for adding the end pixel extracted by the end pixel extractor as an edge pixel based on a pixel value of a pixel located within a first range from the end pixel, a straight line extractor for extracting straight lines from the edge pixel extracted by the edge pixel extractor and the end pixel added as an edge pixel by the edge pixel addition module, and a rectangle region detector for detecting a rectangle region from the extracted straight lines.