Character Offset Detection via Curve Fitting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for text image analysis primarily focus on detecting text lines and do not effectively analyze character offsets, which are crucial for authenticating seal or rubbing images, especially in terms of font size and location consistency.
Innovation Solution
A character offset detection method and system that involves acquiring a text image, performing character separation, calculating center points, determining an optimal fitted curve, and analyzing character offsets based on this curve to obtain detection results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If text line detection methods are used, then text location can be identified, but character offset analysis and authentication capability are lost
Solution Approach 1:
The patent segments the text detection task into two independent stages: first detecting text line locations using existing methods, then analyzing character offsets within each text line region. This segmentation allows the system to build character offset detection capability without completely redesigning the entire text detection system, thus improving measurement precision while controlling device complexity.
Solution Approach 2:
The patent transitions from one-dimensional text line detection to two-dimensional character offset analysis by introducing a fitted curve dimension. The fitted curve represents the expected character position trajectory, and deviations from this curve in multiple dimensions enable precise character offset detection, thereby resolving the contradiction between detection accuracy and system complexity.
2Reliability
If character analysis is performed, then authentication capability is improved, but detection complexity and computational cost increase
Solution Approach 1:
The patent performs preliminary actions by first detecting text line locations and extracting character bounding boxes before conducting the complex character offset analysis. This preliminary segmentation reduces the computational complexity of subsequent authentication analysis by focusing calculations only on relevant character regions rather than the entire image, thereby improving reliability while controlling complexity.
Solution Approach 2:
The patent introduces a fitted curve as an intermediary representation that bridges the gap between simple text line detection and complex character offset analysis. The fitted curve serves as a mediator that encodes expected character position patterns, enabling reliable authentication through deviation analysis without requiring direct complex comparisons between all character pairs, thus reducing computational complexity.
3Measurement precision
If existing text detection methods are used, then text location is detected, but character position consistency and font size analysis are unavailable
Solution Approach 1:
The patent applies curvature by introducing a fitted curve that models the expected trajectory of character positions. Instead of using straight lines or simple geometric shapes, the fitted curve captures the natural curvature and flow of character arrangements in the text. This curvature-based approach enables precise measurement of character position deviations while simplifying the measurement process by providing an intuitive reference framework, thus resolving the contradiction between measurement precision and measurement difficulty.
Data Source
AI summary
The present disclosure discloses a character offset detection method and system. The method includes: acquiring a text image; performing character separation based on the text image to obtain a character text region; calculating a center point of each rectangular box in the character text region to obtain a center point set; determining an optimal fitted curve based on the center point set; and analyzing character offset based on the optimal fitted curve to obtain an offset result. The present disclosure realizes detection of the character offset based on curve fitting, so that the accuracy of detection is improved.