Document Layout Detection Using Blockwise Black Point Counts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing techniques fail to accurately determine the layout of multiple images on a single page and often misidentify division positions, and they also struggle to detect and delete blank-sheet images in divided images.
Innovation Solution
An image processing apparatus that divides a single page into blocks, counts black points in each block, and uses these counts to determine if the document is a layout image and detect and delete blank-sheet images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If feature values of lines are extracted to determine layout images, then determination can be made based on blank area distribution, but division position may be erroneously determined when non-feature pixels are distributed close to one another
Solution Approach 1:
The patent divides the image into multiple blocks and calculates black point counts for each block individually. This segmentation approach allows the system to analyze local density characteristics rather than relying on global feature values, thereby accurately identifying division positions even when non-feature pixels are closely distributed.
Solution Approach 2:
The patent replaces the mechanical feature extraction method with a statistical counting method. Instead of extracting complex line feature values and analyzing blank area distributions, the system simply counts black points in each block, substituting a simpler statistical approach that proves more reliable for determining division positions.
2Adaptability or versatility
If feature value extraction method is used, then layout determination is possible, but detection and deletion of blank-sheet images in divided images cannot be achieved
Solution Approach 1:
The patent replaces the feature value extraction mechanism with a black point counting mechanism. This substitution enables the system to not only determine layout images but also to detect and delete blank-sheet images by analyzing the distribution and density of black points across divided blocks, achieving a functionality that the original feature-based method could not provide.
3Reliability
If black point counting method is used instead of feature value extraction, then blank-sheet image detection and deletion becomes possible, but additional processing steps are required
Solution Approach 1:
The patent segments the image into blocks and performs black point counting on each block. While this adds a division step, the counting operation itself is computationally simpler than feature value extraction and analysis. The segmentation enables parallel processing and facilitates blank-sheet detection through straightforward density comparison across blocks.
Solution Approach 2:
The patent changes the measurement parameter from complex feature values to simple black point counts. This parameter transformation simplifies the underlying computation while enabling new capabilities. The system uses thresholds on black point counts to identify blank sheets, replacing complex feature analysis with simpler statistical comparisons.
Data Source
AI summary
An image processing apparatus includes a processor configured to: obtain image information by reading a document image; divide a single page in the obtained image information on the basis of a predetermined block unit; count black points in each block obtained through the division; use the counting result of black points in each block to determine whether the document image is a layout image in which multiple images are allocated in a single page, and to detect and delete a blank-sheet image in the layout image.