Differential OCR Handwriting Scoring for Objective Motor Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current motor skill evaluations, particularly for conditions like ALS and PD, are subjective and inconsistent, relying on human assessment and language proficiency, leading to inaccuracies and self-bias issues.
Innovation Solution
An objective method using two image-to-text models with varying performance levels to analyze handwritten images, generating a handwriting skill score by comparing their outputs, which can be used to track motor skill deterioration over time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If subjective human assessment is used for motor skill evaluation, then language proficiency and human judgment are utilized, but consistency and objectivity deteriorate due to self-bias and subjectivity
Solution Approach 1:
The patent replaces the mechanical system of human visual assessment and language interpretation with an optical character recognition (OCR) system that processes handwritten images directly. This substitution eliminates human subjectivity and self-bias by using automated image-to-text conversion models to evaluate handwriting characteristics objectively, thereby improving evaluation consistency while maintaining ease of operation through digital processing.
2Productivity
If single model analysis is used for handwriting evaluation, then processing speed is maintained, but measurement precision deteriorates due to lack of differential assessment
Solution Approach 1:
The patent segments the handwriting evaluation process into multiple independent analysis stages using different OCR models with varying performance levels. By dividing the assessment into differential comparisons between multiple model outputs rather than relying on a single model, the system achieves more precise measurement of handwriting skill while maintaining processing efficiency through parallel model execution.
3Reliability
If automated image-to-text models are used for handwriting analysis, then objectivity is improved, but device complexity increases due to multiple model requirements
Solution Approach 1:
The patent implements a universal evaluation framework where multiple OCR models with different performance levels are integrated into a single differential assessment system. This multi-functional approach allows the same system architecture to accommodate various model capabilities, achieving high evaluation objectivity while managing complexity through a unified processing pipeline that handles multiple model outputs systematically.
Data Source
AI summary
An embodiment analyzes, using a first image-to-text model, a handwritten image, the analyzing resulting in a first text output corresponding to the handwritten image. An embodiment analyzes, using a second image-to-text model with a higher performance level than the first image-to-text model, the handwritten image, the analyzing resulting in a second text output corresponding to the handwritten image. An embodiment generates, by analyzing a difference between the first text output and the second text output, a handwriting skill score.