Document Image Rule Extraction Using Multi-Directional Pixel Scanning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods struggle to accurately extract rules from document images while maintaining high processing speed, often failing to recognize oblique lines and taking too long to extract rules, leading to compromised readability and inefficient compression.
Innovation Solution
An image processing apparatus and program that scans image data in multiple directions to detect series of pixels as rules, with adjustable thresholds for precise extraction, including oblique lines, to ensure accurate rule detection and efficient compression.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If rule extraction is performed using existing methods, then processing speed is improved, but measurement precision of rule detection deteriorates
Solution Approach 1:
The patent segments the rule detection process into two distinct phases: first detecting long pixel series with a relaxed threshold to capture potential rules including oblique lines, then detecting shorter pixel series with a stricter threshold to refine rule identification. This segmentation allows the system to maintain high processing speed while improving detection precision by handling different rule characteristics in separate stages.
Solution Approach 2:
The patent performs preliminary detection of long pixel series before detecting shorter series. By first identifying potential rule regions using a lower threshold, the system prepares the groundwork for more precise detection in the second phase. This preliminary action ensures that oblique lines and other complex rule structures are not missed, while maintaining efficient processing through staged detection.
2Device complexity
If uniform resolution compression is applied to the whole image, then device complexity is reduced, but manufacturing precision of character readability deteriorates
Solution Approach 1:
The patent applies different compression resolutions to different regions of the image based on their content. Character regions are compressed with high resolution to preserve readability, while non-character regions use lower resolution compression. This local quality approach maintains character readability without requiring uniform high-resolution processing of the entire image, thus balancing complexity and precision.
Solution Approach 2:
The patent performs preliminary rule extraction and character region identification before compression. By pre-segmenting the image into character and non-character regions based on detected rules, the system prepares the groundwork for differential compression. This preliminary action enables the compression process to apply appropriate resolution to each region, ensuring character readability while managing overall complexity.
3Manufacturing precision
If high resolution compression is applied to ensure character readability, then manufacturing precision of character readability is improved, but loss of substance in terms of data size increases
Solution Approach 1:
The patent applies high resolution compression only to character regions where readability is critical, while applying lower resolution compression to non-character regions. This local quality approach ensures that characters maintain their readability with high resolution, while reducing overall data size by compressing non-essential areas at lower resolution. The differential compression strategy optimizes the balance between readability and data size.
Solution Approach 2:
The patent applies high resolution compression selectively to only the portions of the image containing characters, rather than uniformly to the entire image. This partial action approach maintains character readability where needed while avoiding the excessive data size that would result from applying high resolution compression to the whole image, including non-character regions.
4Productivity
If existing rule extraction methods are used, then processing speed is maintained, but reliability of rule extraction deteriorates due to failure to recognize oblique lines
Solution Approach 1:
The patent segments the rule extraction process into two phases: first detecting long pixel series with a relaxed threshold to capture oblique lines and potential rules, then detecting shorter series with a stricter threshold to refine detection. This segmentation improves reliability by ensuring oblique lines are not missed, while maintaining processing speed through efficient staged detection rather than using a single complex algorithm.
Solution Approach 2:
The patent changes the detection threshold parameter between two detection phases. The first phase uses a lower threshold to capture potential rules including oblique lines, while the second phase uses a higher threshold for refinement. This parameter change strategy improves reliability by adapting the detection sensitivity to different stages of the process, ensuring that oblique lines and other complex rule structures are reliably detected without sacrificing processing speed.
Data Source
AI summary
An image processing apparatus scans image data along a row in an x direction, and if a series of black pixels having a length equal to or more than a first threshold is detected, the series of black pixels is determined as a rule and erased. Furthermore, if a series of black pixels having a length equal to or more than a second threshold that is connected to the rule and extends in a y direction is detected, the series of black pixels is determined as a rule and erased.