Handwriting Recognition System with Spatial Learning Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing information processing systems require significant user effort for checking and modifying hand-written data, increasing the user burden.
Innovation Solution
An information processing apparatus and system that automatically corrects various types of information obtained through writing operations by acquiring learning information on spatial or semantic relations between objects, performing recognition processes on stroke data, and adjusting the position or size of objects based on this information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If automatic correction is implemented, then user burden is reduced, but device complexity increases
Solution Approach 1:
The system performs preliminary learning of spatial relationships between objects during a training phase, storing correction rules in advance. When handwriting is input, the system applies pre-established correction logic based on learned patterns, reducing real-time processing complexity while maintaining automatic correction capability.
Solution Approach 2:
The system introduces a learning module as an intermediary between data input and correction execution. This module pre-processes spatial relationship data and generates correction rules that are then applied by the correction unit, separating the complex learning process from the simpler application process.
2Measurement precision
If learning information is acquired and processed, then correction accuracy is improved, but processing time increases
Solution Approach 1:
The system performs learning and generates correction rules in advance, before actual handwriting correction is needed. This pre-processing approach allows the system to store learned spatial relationships and correction patterns, enabling fast application during actual use without repeated complex processing.
Solution Approach 2:
The correction process is divided into distinct phases: a learning phase where spatial relationships are analyzed and stored, and an execution phase where pre-established rules are applied. This segmentation allows complex learning to occur separately from time-sensitive correction operations.
Data Source
AI summary
Provided is an information processing apparatus including a processor and a memory storing instructions that, when executed by the processor, cause the information processing apparatus to: acquire learning information representing a result of learning on a spatial relation between adjacent objects, perform a recognition process on stroke data representing a collection of strokes to recognize a first object and a second object that has been inputted via a writing input after the first object, and perform an adjustment process on the stroke data based on the learning information such that a position or size of the second object is adjusted with the first object being fixed.