Logogram Prediction Using Sub-Logogram Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face challenges in accurately predicting logograms from cursive written inputs due to connective strokes that impede recognition, leading to suboptimal logogram prediction.
Innovation Solution
An apparatus and method that utilize a processor and memory to generate logogram predictions by identifying connective points, removing connective strokes, and forming sub-logograms, which are then used to predict the target logogram, improving prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If connective strokes are retained in cursive written input, then the input can be written continuously and efficiently, but the recognition accuracy of logograms deteriorates
Solution Approach 1:
The system segments the cursive written input by identifying connective points and separating connective strokes from the main logogram structure. This segmentation allows the input to be divided into sub-logograms that can be independently recognized and processed, resolving the conflict between continuous writing and accurate recognition.
Solution Approach 2:
The system extracts and removes connective strokes from the cursive input by identifying them based on stroke direction, curvature, and position relative to the main logogram. By taking out these harmful connective elements, the system preserves writing efficiency while improving recognition accuracy of the underlying logogram.
2Measurement precision
If connective strokes are removed to improve recognition, then logogram prediction accuracy improves, but the complexity of the processing system increases
Solution Approach 1:
The system performs preliminary identification of connective points and connective strokes before final logogram recognition. By pre-processing the input to mark and separate connective elements, the system simplifies the subsequent recognition process and reduces overall computational complexity despite the additional initial processing step.
Solution Approach 2:
The system introduces intermediate structures such as sub-logograms and connective point markers that facilitate the transition from cursive input to recognized logogram. These intermediary representations simplify the complex task of direct recognition by breaking it into manageable stages with clear intermediate goals.
Data Source
AI summary
For predicting a target logogram, code generates a logogram prediction from a contiguous written input. In addition, the code removes a connective stroke between two connective points in the contiguous written input. The code generates a sub-logogram prediction for a sub-logogram formed by removing the connective stroke. In addition the code predicts a target logogram for the contiguous written input based on both the logogram prediction and the sub-logogram prediction.