Handwriting Recognition Using Spatial Syntactic and Semantic Rules
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current technologies lack efficient solutions for accurately recognizing both math and text content in handwriting, particularly in mixed content scenarios.
Innovation Solution
A method implemented by a computing device for processing math and text in handwriting, involving a three-level analysis: symbol classification based on predefined recognition rules, spatial syntactic classification to assess spatial relationships between symbols, and semantic classification to establish connections between symbols, thereby distinguishing math from text content.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional text classifiers are used for handwriting recognition, then text content can be recognised, but math content cannot be accurately distinguished from text
Solution Approach 1:
The patent segments the handwriting recognition task into three distinct classification levels: symbol-level classification (individual character/math symbol), word-level classification (sequences of symbols), and document-level classification (overall document type). This segmentation allows the system to handle different content types with appropriate classification strategies at each level, resolving the contradiction between accuracy for specific content types and versatility across multiple content types.
Solution Approach 2:
The patent creates a universal classification framework that can handle both text and math content through a single system. The multi-level classification architecture uses the same basic structure (symbol → word → document) for both text and math content, making the system versatile while maintaining high accuracy for each content type through content-specific classification rules at each level.
2Measurement precision
If a single classification approach is used for all handwriting symbols, then the system is simple, but it cannot accurately distinguish math symbols from text symbols
Solution Approach 1:
The patent adds a vertical dimension to the classification system by introducing multiple classification levels (symbol-level, word-level, document-level) rather than using a single horizontal classification approach. This dimensional expansion allows the system to capture different aspects of handwriting content at each level, improving classification accuracy while organizing complexity in a structured, manageable way.
Solution Approach 2:
The classification system is designed to be dynamic, adapting the classification depth and methods based on the content being analyzed. The system can operate at different levels depending on the context, and the classification rules are applied dynamically based on the detected content type, allowing high precision without requiring all possible complexity to be active simultaneously.
3Productivity
If handwriting recognition system is designed for text only, then text recognition is accurate, but it fails to recognise math content
Solution Approach 1:
The patent applies preliminary classification at the symbol level before proceeding to word-level and document-level classification. This preliminary action quickly identifies math-specific symbols and structures, allowing the system to route content appropriately and maintain high recognition speed while ensuring math content is accurately identified through early detection of math-specific features.
Data Source
AI summary
The invention relates to a method implemented by a computing device for processing math and text in handwriting, comprising: identifying symbols by performing handwriting recognition on a plurality of strokes; classifying, as a first classification, first symbols as either a text symbol candidate or a math symbol candidate with a confidence score reaching a first threshold; classifying, as a second classification, second symbols other than first symbols as either a text symbol candidate or a math symbol candidate with a respective confidence score by applying predefined spatial syntactic rules; updating or confirming, as a third classification, a result of the second classification by establishing semantic connections between symbols and comparing the semantic connections with the result of the second classification; and recognising each symbol as either text or math based on a result of said third classification.