Handwritten text processing method, apparatus, device, and readable storage medium

By activating the feature vectors of handwritten characters and answer characters, the accuracy problem of optical character recognition algorithms under the difference of users' handwriting habits is solved, realizing efficient recognition and correctness judgment of handwritten text, and improving the learning assistance effect of learning machines and other terminals.

CN122290146APending Publication Date: 2026-06-26BEIJING XUEDIRUANJIAN DEVELOPMENT CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
BEIJING XUEDIRUANJIAN DEVELOPMENT CO LTD
Filing Date
2026-03-26
Publication Date
2026-06-26

AI Technical Summary

Technical Problem

Existing optical character recognition algorithms struggle to accurately recognize handwritten text when faced with significant differences in users' handwriting habits. This leads to a decrease in the accuracy of judging the correctness of handwritten text, misleads users into perpetuating incorrect writing habits, and negatively impacts language learning outcomes.

Method used

By constructing a nonlinear feature mapping mechanism based on the range of activated feature segments, the feature vectors of handwritten characters and answer characters are activated, weak features are reduced and strong features are amplified, interference caused by noise and differences in writing habits is dynamically suppressed, and the recognition accuracy is improved by using a dual verification mechanism.

Benefits of technology

It significantly improves the accuracy of judging the correctness of handwritten text, and can effectively deal with situations such as connected strokes, messy character shapes, irregular character spacing, missing or redundant strokes when users write, providing reliable practice feedback and improving language learning efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122290146A_ABST
    Figure CN122290146A_ABST
Patent Text Reader

Abstract

This application provides a handwritten text processing method, apparatus, device, and readable storage medium. By constructing a nonlinear feature mapping mechanism based on the range of activated feature segments, the first character feature vector of the handwritten character and the second character feature vector of the answer character in the character pair aligned in position in the handwritten text and the answer text are subjected to vector activation processing. Then, in the character comparison stage, the original character feature vector and the activated feature vector are fused to perform double verification of the handwritten character and the answer character in the character pair. This can effectively cope with handwritten character recognition and handwritten text correctness judgment under the conditions of serious stroke connection, messy character shape, irregular character spacing, missing or redundant strokes, etc., when the user writes, thus improving the accuracy of handwritten text correctness judgment.
Need to check novelty before this filing date? Find Prior Art