Adaptive Binarization for Document Image Defect Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Images captured from mobile devices often contain defects such as lighting issues and shadows, making it difficult to extract data accurately, especially for larger documents, which increases processing time and power requirements, making mobile document capture inefficient.
Innovation Solution
A computer-implemented method that separates document images into regions of darkness and light, using adaptive binarization and bit plane slicing to identify and remove defects, allowing for targeted filtering and data extraction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If mobile devices are used to capture larger documents, then document capture capability is improved, but processing time and power requirements increase
Solution Approach 1:
The patent divides the document image into multiple bit planes (e.g., 8 bit planes for an 8-bit grayscale image). Each bit plane represents a different range of gray levels and can be processed independently for defect detection and removal. This segmentation allows parallel processing of different bit planes, significantly reducing overall processing time while maintaining the ability to handle large documents.
Solution Approach 2:
The patent applies defect removal processing selectively to only those bit planes that contain defects, rather than processing all bit planes uniformly. By identifying which bit planes require defect removal and applying processing only to those specific planes, the system reduces unnecessary computational overhead and processing time while still achieving effective defect removal where needed.
2Adaptability or versatility
If mobile devices are used to capture larger documents, then document capture capability is improved, but power consumption increases
Solution Approach 1:
The patent segments the image processing task into independent bit plane operations. Each bit plane can be processed separately and potentially in parallel, allowing the system to optimize power usage by activating only the necessary processing units for each bit plane. This reduces overall power consumption compared to processing the entire image as a single unit.
Solution Approach 2:
The patent performs defect removal only on bit planes that contain defects, avoiding unnecessary processing of clean bit planes. This selective approach reduces the total computational workload and corresponding power consumption while maintaining effective defect removal capability for documents of any size.
3Reliability
If traditional defect removal methods are used on larger documents, then defect removal is achieved, but data extraction accuracy decreases due to increased processing complexity
Solution Approach 1:
The patent processes each bit plane independently to detect and remove defects. This segmentation ensures that defect removal operations are applied precisely where needed without affecting other parts of the image. The independent processing of each bit plane maintains the integrity of the original image data while effectively removing defects, thereby preserving data extraction accuracy.
Solution Approach 2:
The patent uses bit planes as intermediary representations of the image data. By transforming the image into multiple bit planes, processing, and then reconstructing the final image, the system achieves precise defect removal while maintaining the overall image quality and data extraction accuracy. The bit planes serve as intermediaries that facilitate accurate defect removal without degrading the final image quality.
Data Source
AI summary
A method is described to increase the efficiency of the removal of defects from document images by reorienting the conceptual framework within which an image is filtered. Rather than arbitrarily applying a filter to an entire landscape of a document image, the disclosure describes a methodology by which a document image is separated into regions of darkness and regions of light, or viewed alternatively, regions of darkness and regions of lack of darkness. Filtering is then adaptively applied to each region to remove defects.