Underline Removal for Character Recognition Integrity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing document image processing methods face inefficiencies and accuracy issues due to the presence of non-text objects like charts, tables, and underlines, which hinder character recognition and conversion processes, particularly in large document repositories.
Innovation Solution
A method and apparatus for removing underlines from images by identifying and excluding rows of foreground pixels that do not meet a minimum length requirement, generating a reduced underline, and restoring intersection areas to maintain character integrity, thereby creating a de-underlined connected component.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If underline removal is performed by completely removing the underline, then the underline interference is eliminated, but character integrity is damaged
Solution Approach 1:
The underline is segmented into two parts: the portion that intersects with characters (intersection area) and the portion that does not intersect (non-intersection area). Only the non-intersection area is removed, while the intersection area is preserved to maintain character integrity. This segmentation allows selective removal of harmful underline portions while preserving beneficial portions that support character structure.
Solution Approach 2:
Different treatment is applied to different parts of the underline based on their spatial relationship with characters. The intersection area (where underline meets character) is preserved with original properties, while the non-intersection area is removed. This local differentiation ensures that character integrity is maintained while eliminating underline interference in clear regions.
2Productivity
If traditional underline removal methods are used, then processing simplicity is maintained, but computational efficiency is reduced
Solution Approach 1:
The method performs preliminary identification of the intersection area between the underline and character before executing the removal operation. By pre-calculating and marking the intersection area, the system avoids unnecessary computations during the actual removal phase, thereby improving overall processing efficiency and reducing computational time.
3Reliability
If the entire underline is removed, then underline interference is completely eliminated, but character structure is compromised
Solution Approach 1:
The underline is divided into intersection area (overlapping with character) and non-intersection area (clear of character). The segmentation enables selective removal of only the non-intersection portion, preserving the character's structural integrity while eliminating interference from the underline in clear regions.
Solution Approach 2:
The solution applies different operations to different spatial zones: preservation in the intersection area and removal in the non-intersection area. This localized approach maintains character structure where needed while eliminating harmful underline effects where safe.
Data Source
AI summary
Complete removal of an underline which intersects a character may cause problems in a subsequent character recognition or conversion process, when parts of the character which coincided with the underline are also removed. To help reduce the problems, parts of underline may be removed from an image while parts of the character that coincide with the underline are maintained in the image. Areas where the character coincides with the underline are defined from a reduced version of the underline. When the underline is removed, the areas where the character coincide with the underline are maintained in a second image. The second image may then be subjected to a character recognition or conversion process with potentially fewer problems.