Personalized Handwriting Recognition via Model Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing handwritten character recognition systems use a common trained model for all users, which does not account for individual writing idiosyncrasies, leading to suboptimal recognition rates.
Innovation Solution
An image processing apparatus and method that selects a trained model specific to each user from a plurality of models based on their unique writing characteristics, allowing for personalized character recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a common trained model is used for all users, then the device complexity is reduced and ease of operation is improved, but the recognition accuracy deteriorates due to inability to account for individual writing idiosyncrasies
Solution Approach 1:
The system performs preliminary actions by pre-training multiple recognition models with different training data corresponding to different users before actual use. When a sheet is read, the system selects and uses the pre-prepared model that matches the user's writing characteristics, thereby achieving high recognition accuracy without complex real-time adaptation.
Solution Approach 2:
The system changes the parameter of model selection based on user identification. Different users are associated with different trained models, and the system selects the appropriate model by changing which model is activated. This allows the recognition system to adapt to individual writing styles by selecting models with different training parameters.
2Measurement precision
If individually trained models are used for each user, then recognition accuracy is improved, but the device complexity increases due to managing multiple models
Solution Approach 1:
Multiple recognition models are trained in advance with different training data corresponding to different users. The system prepares these models beforehand so that when a sheet is read, it can simply select the appropriate pre-trained model based on user identification, avoiding the need for complex real-time model training or adaptation.
Solution Approach 2:
The system creates a universal solution by training multiple models with different training data that can handle different users' writing styles. Each model serves a specific user, but collectively they provide a universal recognition system that adapts to any user in the set, achieving high accuracy across multiple users without requiring real-time customization.
3Adaptability or versatility
If multiple trained models are maintained for different users, then adaptability to individual writing styles is improved, but the loss of time increases due to model selection process
Solution Approach 1:
The system performs model selection based on user identification information extracted from the read sheet. By preliminarily associating users with specific models and using quick identification methods, the system minimizes selection time while maintaining adaptability to individual writing styles.
Data Source
AI summary
An image processing apparatus obtains image data generated by reading a sheet on which handwritten characters are written. The image processing apparatus selects a trained model to be used for character recognitions of the handwritten characters on the sheet from a plurality of trained models and executes character recognitions by using the selected trained model.