Patterned Artifact Removal in Bitonal Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for removing patterned artifacts from bitonal images often erode desired information, leading to failures in optical character recognition (OCR) and other processes, as they fail to distinguish between artifacts and pixels of interest.
Innovation Solution
A computer-implemented system and method that identifies patterned artifacts in bitonal images using erosion algorithms based on stroke width, defines artifact boundaries, applies filters, and performs binary operations to remove unwanted artifacts while preserving desired information, including text identification and cleaning steps to ensure accurate removal.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If erosion methods are used to remove patterned artifacts, then the unwanted patterned artifact is minimized, but the pixels/items of interest are also eroded causing failures in OCR
Solution Approach 1:
The patent applies different processing treatments to different regions of the image. It identifies artifact regions versus text regions and applies erosion selectively to artifact regions while preserving text regions. This is achieved through detecting continuous black pixel runs and applying region-specific morphological operations, ensuring that artifacts are removed while text integrity is maintained.
Solution Approach 2:
The patent segments the image into different regions based on the presence of artifacts versus text content. By detecting continuous black pixel runs and analyzing image characteristics, it divides the image into artifact regions and text regions, then applies different processing strategies to each segment. This segmentation allows selective artifact removal without affecting text areas.
2Object-affected harmful factors
If aggressive artifact removal is applied, then artifact reduction is improved, but OCR accuracy deteriorates due to erosion of text pixels
Solution Approach 1:
The patent implements local quality by applying different processing intensities to different image regions. Artifact regions undergo aggressive erosion while text regions are protected. The system detects continuous black pixel runs characteristic of artifacts and applies region-specific morphological operations, ensuring high artifact removal efficiency in artifact areas while maintaining OCR accuracy in text areas.
Solution Approach 2:
The patent introduces an intermediary detection mechanism that identifies artifact regions before applying removal operations. By detecting continuous black pixel runs and analyzing image characteristics, it creates a mask or region definition that mediates between artifact removal and text preservation. This intermediary step ensures that subsequent erosion operations are applied only where needed, preventing collateral damage to text.
3Object-affected harmful factors
If erosion algorithms are applied to remove artifacts, then the unwanted pattern is minimized, but the integrity of desired information is compromised
Solution Approach 1:
The patent applies local quality by treating artifact regions and text regions differently. It detects continuous black pixel runs and applies erosion selectively to artifact regions while preserving text regions. This ensures that the unwanted pattern is minimized in artifact areas while the integrity of desired information in text areas is maintained.
Solution Approach 2:
The patent segments the image into artifact regions and text regions based on detection of continuous black pixel runs and image characteristics. By dividing the image into these segments, it can apply erosion to artifact segments while protecting text segments, thus removing the unwanted pattern without compromising the integrity of desired information.
Data Source
AI summary
Images in bitonal formats often include watermarks, stamps, or other patterns and artifacts. These patterned artifacts may be represented as a series of geometric points, dots, and/or dashes in the general shape of the original pattern. These patterned artifacts make other processes such as optical character recognition (OCR) difficult or impossible when items or pixels of interest are also found within the pattern of such artifact(s). Current patterned artifact removal solutions use methods of erosion to minimize the unwanted patterned artifact. However, such methods also erode the pixels/items of interest which, in turn, cause failures in other processes, such as OCR, that are desired to be carried out on or with the image.