Character Normalization via Adaptive Margin Bounding Boxes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional normalization methods for characters in images, such as adding margins around bounding boxes, often distort character shapes or include excessive background information, leading to lowered recognition rates, especially for characters with narrow widths and varying font types.
Innovation Solution
A method that determines and adds a margin around a bounding box based on information from adjacent or related characters, using mean or median values of character widths to normalize the bounding box area, ensuring accurate recognition of characters with narrow widths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a constant margin is added around all bounding boxes, then normalization is simplified and processing speed increases, but recognition accuracy of narrow characters deteriorates because they remain one-colored
Solution Approach 1:
The patent applies different margin values to different characters based on their individual width characteristics. Narrow characters receive larger margins while wide characters receive smaller margins, creating localized quality adjustments that preserve recognition accuracy for each character type while maintaining efficient processing
Solution Approach 2:
The patent dynamically changes the margin parameter based on character width measurements. By calculating the width of each character and adjusting the margin accordingly (larger margins for narrow characters, smaller margins for wide characters), the system adapts the normalization parameters to achieve both speed and accuracy
2Manufacturing precision
If margins are added based on character width to improve narrow character recognition, then recognition accuracy improves, but device complexity and processing time increase
Solution Approach 1:
The system performs self-service by automatically measuring each character's width and determining its own appropriate margin value without requiring external intervention or complex configuration. The normalization process serves itself by using the character's own geometric properties to guide the margin addition
Solution Approach 2:
The patent replaces complex mechanical or manual normalization adjustments with an automated computational system that calculates margins based on character width measurements, substituting a simple algorithmic approach for what would otherwise require complex manual intervention
3Ease of operation
If stretching method is used to normalize bounding box sizes, then all characters can be made equal size for processing, but character shape distortion occurs
Solution Approach 1:
Instead of applying uniform stretching to all characters, the patent applies localized quality preservation by adding margins that maintain each character's original aspect ratio and shape characteristics, allowing different characters to retain their individual shape qualities while achieving uniform processing dimensions
Data Source
AI summary
A method for normalizing at least one character included in an image is provided. The method includes steps of: (a) an apparatus for normalizing a specific character adding or supporting to add at least one margin, which is determined by referring to information on at least one another character related to the specific character detected in the image, around a bounding box which includes the specific character; and (b) the apparatus normalizing or supporting to normalize an area of the bounding box around which the at least one margin is added.