Document Image Blank Sheet Detection Using CIELAB a*b* Histogram
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image processing systems inaccurately determine whether a document is blank by relying on RGB color space histograms, leading to misclassification of non-blank documents as blank when all RGB maximum values and differences are the same, regardless of content differences from the background.
Innovation Solution
An image processing system that reads document images in the CIELAB color space, generates an a*b*-plane two-dimensional histogram, counts black spots in each bin, subtracts background-related black spots, and determines if the result exceeds a threshold to identify non-blank sheets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If histograms of maximum values and maximum differences of RGB are used to determine blank sheets, then the determination process is simple, but the accuracy deteriorates when all RGB maximum values and differences are the same even if content differs from background
Solution Approach 1:
The patent changes the color space parameter from RGB to CIELAB, specifically utilizing the a* and b* channels. This parameter change allows the system to distinguish between background and content based on color differences in the a*b* plane, resolving the issue where RGB maximum values are identical despite content differences. The histogram is now created in the CIELAB color space rather than RGB, enabling accurate detection of non-blank sheets through color variation analysis.
2Ease of manufacture
If RGB color space histograms are used, then the processing method is straightforward, but false positives occur when content color matches background luminance characteristics
Solution Approach 1:
The patent introduces the CIELAB color space as an intermediary between the raw image data and the blank sheet determination. By transforming RGB values to CIELAB and creating histograms in the a*b* plane, the system uses this intermediate color space representation to accurately distinguish content from background. This intermediary approach maintains processing simplicity while significantly improving determination reliability by avoiding the luminance-based false positives inherent in direct RGB analysis.
3Use of energy by moving object
If maximum RGB values are compared for blank sheet detection, then computational load is low, but accuracy decreases when documents have uniform color characteristics
Solution Approach 1:
The patent transitions from analyzing only the maximum values in RGB space to creating a two-dimensional histogram in the CIELAB a*b* plane. This dimensional change allows the system to capture color distribution patterns across both a* and b* channels simultaneously. The 2D histogram approach provides richer information about color variations in the document, enabling accurate detection of non-blank sheets even when overall luminance or single-channel maximum values are uniform, while maintaining reasonable computational efficiency.
Data Source
AI summary
An image processing system includes an image reader that reads a document image and that generates image information regarding a CIELAB color space, and a processor configured to: obtain an a*b*-plane two-dimensional histogram for the image information; obtain a background color of the document image using the a*b*-plane two-dimensional histogram and count a number of black spots of the image information for each of bins of the a*b*-plane two-dimensional histogram; subtract the number of black spots in a bin determined to include the background color from a total number of black spots in all the bins; and determine, if an obtained value is larger than a predetermined threshold, that the document image indicates a non-blank sheet.