Binary Image Blob Comparison Using Dilation and Cross-Check
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image compression techniques for scanned documents suffer from insufficient compression rates and often compromise the quality of the reconstructed image, particularly in terms of legibility and rendering, especially when dealing with documents containing text elements of different colors and backgrounds.
Innovation Solution
A method that segments digital images into foreground, background, and binary mask layers, using a symbol-based compression technique to minimize errors in symbol comparison, and adjustable resolution reduction factors to balance compression quality, employing techniques like dilation and threshold checks to ensure accurate matching and reduce erroneous matches.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of substance
If existing image compression techniques are applied to scanned documents, then compression is achieved, but the compression rate is insufficient and the quality of the reconstructed image deteriorates
Solution Approach 1:
The patent segments the binary image into multiple equivalence classes by comparing symbols and grouping identical or similar characters together. This segmentation allows the compression algorithm to represent repeated symbols more efficiently, achieving higher compression rates while preserving image quality through accurate symbol reconstruction.
2Productivity
If symbol comparison methods are used to classify symbols into equivalence classes, then compression efficiency improves, but erroneous matches occur reducing reliability
Solution Approach 1:
The patent employs a feedback mechanism in the symbol comparison process where the dilation tolerance is adjusted based on the comparison results. The algorithm checks if dilated symbols match and uses this feedback to refine the equivalence class grouping, thereby reducing erroneous matches while maintaining compression efficiency.
Solution Approach 2:
The patent changes the parameter of dilation tolerance dynamically during symbol comparison. By adjusting the dilation level and applying threshold checks, the system optimizes the balance between matching sensitivity and accuracy, reducing false positives while maintaining high compression efficiency.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention is directed to a method for comparing in a binary image a first pixel blob with a second pixel blob to determine whether they represent matching symbols. The method comprises the steps of dilating the first blob and checking if the second blob fits inside the dilated first blob, and dilating the second blob and checking if the first blob fits inside the dilated second blob. If the dilating steps result in a match, the method checks whether the match is erroneous or not by checking if the non-dilated first pixel blob contains a line not present in the non-dilated second pixel blob and vice versa. This latter step further comprises directly comparing corresponding sets of NxN crosses of two crossing lines in the non-dilated first and second pixel blobs, wherein each corresponding set of NxN crosses comprises a first NxN cross considered on the non-dilated first pixel blob and a second NxN cross considered on the non-dilated second pixel blob in the same position, and declaring an erroneous match if for at least one of the sets one of the two crossing lines of the first NxN cross has a different color from the other line of the two crossing lines of the second NxN cross, wherein N is a number of bits.