Character Region Extraction via Stroke Width Calculation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing character region detection methods in images face challenges in accurately calculating character stroke widths due to errors caused by non-perpendicular scan lines and character inclinations, leading to incorrect detection.
Innovation Solution
A system and method that calculates correct character stroke widths by producing a binary image, extracting character outlines, setting representative stroke widths and angles, and determining character regions based on the ratio of effective stroke widths and angles compared to the entire outline.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If character stroke widths are calculated along a scan line in an image region, then the detection process is simple, but the stroke width calculation becomes inaccurate when the scan line does not pass perpendicularly through character strokes or when characters are inclined
Solution Approach 1:
The patent transitions from one-dimensional scan line-based stroke width calculation to two-dimensional character outline-based calculation. By extracting the character outline and calculating stroke widths along the outline perimeter rather than along fixed horizontal scan lines, the method adapts to character orientation and inclination, thereby improving measurement precision while maintaining operational simplicity through automated outline extraction.
Solution Approach 2:
The patent introduces dynamic adaptation by calculating stroke widths along the character outline rather than along fixed scan lines. The calculation direction dynamically follows the character's actual geometry and orientation, allowing accurate stroke width measurement regardless of character inclination or rotation, thus resolving the precision issue without complicating the overall detection process.
2Loss of information
If character regions are detected based on edge, color and contrast characteristics, then various character information can be recognized, but incorrect stroke width calculation leads to erroneous character region determination
Solution Approach 1:
The patent implements a feedback mechanism where character outline information is used to verify and refine stroke width calculations. By comparing the calculated stroke widths along the outline with expected character properties, the system can confirm whether a detected region truly contains characters, thereby improving reliability without losing character information.
Solution Approach 2:
The patent performs preliminary character outline extraction and stroke width calculation along the outline before final character region determination. This preliminary action ensures that accurate stroke width information is available for subsequent verification, improving the reliability of character region detection while preserving all character information for recognition.
Data Source
AI summary
A character region extracting apparatus and method which extract a character region through the calculation of character stroke widths are provided. The method includes producing a binary image including a candidate character region from an original image; extracting a character outline from the candidate character region; acquires character outline information for the extracted outline; setting a representative character stroke width and a representative character angle in each of the pixels forming the outline, based on the character outline information; and determining a character existing region in the candidate character region by confirming the ratio of effective representative stroke widths and effective angles as compared to the entire length of the outline. Accordingly, it is possible to efficiently determine whether one or more characters exist in the candidate character region.