Document Boundary Detection Using Shadow Verification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Vertical line noise caused by smudges or scanner components can lead to incorrect detection of the boundary line between document and background regions in image data, especially when the brightness difference between the two is small, resulting in errors in pixel classification and boundary line approximation.
Innovation Solution
An image processing apparatus and system that includes a candidate pixel detector, classifier, approximate line calculator, provisional line determiner, shadow detector, and boundary line determiner to accurately identify and correct the boundary line by classifying candidate pixels, calculating approximate lines, detecting shadows, and determining the correct boundary line based on the presence of a shadow image within a predetermined distance from the provisional line.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional boundary detection methods are used, then the detection process is simple, but detection accuracy deteriorates due to vertical line noise and small brightness differences
Solution Approach 1:
The patent segments the boundary detection process into multiple independent modules: candidate pixel detection, coordinate classification, approximate line calculation, provisional line determination, shadow image detection, and final boundary line determination. Each module handles a specific aspect of the detection task, improving accuracy while maintaining manageable system complexity through functional decomposition.
Solution Approach 2:
The patent introduces shadow image detection as an intermediary verification step between provisional line determination and final boundary line determination. The shadow image acts as a mediator to validate whether the detected boundary line position is correct, filtering out false detections caused by vertical line noise or brightness variations.
2Measurement precision
If multiple detection steps are added to improve accuracy, then detection accuracy improves, but processing time increases
Solution Approach 1:
The patent performs preliminary classification of candidate pixels into coordinate groups and calculates approximate lines before final boundary determination. This preliminary processing organizes the data structure and identifies likely boundary regions, reducing the computational burden of subsequent shadow image detection and final determination steps.
Solution Approach 2:
The patent extracts and focuses computational resources on specific regions of interest - particularly areas where shadow images are expected to appear near the boundary. By concentrating processing effort on these critical regions rather than the entire image, the system achieves high accuracy while minimizing overall processing time.
3Measurement precision
If shadow image detection is used to verify boundary lines, then detection accuracy improves, but system complexity increases
Solution Approach 1:
The shadow image detection serves as an intermediary verification mechanism that validates boundary line positions without requiring complex physical modifications to the scanner. The shadow image naturally forms during the scanning process and can be detected through standard image processing, providing verification capability while adding minimal system complexity.
Solution Approach 2:
The patent uses the shadow image as a copy or reflection of the document edge that appears on the backing member. This shadow copy provides independent verification information about the boundary position without requiring direct physical contact or additional sensing elements, thereby verifying accuracy with minimal added complexity.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The system effectively improves the accuracy of boundary line detection by reducing errors caused by vertical line noise and small brightness differences, ensuring precise separation between document and background regions.
Implementation Method 1
a shadow detector for detecting a shadow image of an edge of the document within a predetermined distance from the provisional line
Data Source
AI summary
An image processing apparatus including a candidate pixel detector for detecting candidate pixels of boundary lines of sides of a document region, a classifier for classifying coordinates of the candidate pixels into coordinate groups, an approximate line calculator for calculating approximate lines for the boundary line based on each of the coordinate groups, a provisional line determination unit for determining a provisional line of the boundary line based on the approximate lines that is selected based on the number of candidate pixels that are within a distance from the approximate line, a shadow detector for detecting a shadow image of an edge of the document within a predetermined distance from the provisional line, and a boundary line determination unit for determining whether the boundary line is within the predetermined distance from the provisional line based on the shadow image.