A method for predicting orthokeratology lens parameters based on multi-modal chain tags
By integrating corneal morphology and clinical parameters through a multimodal chain tagging method, and combining static images and dynamic video assessments, the problem of missing high-dimensional features due to dimensional compression in orthokeratology lens parameter prediction was solved, achieving higher accuracy in parameter prediction and better adaptation efficiency.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- THE EYE HOSPITAL OF WENZHOU MEDICAL UNIVERSITY
- Filing Date
- 2026-05-08
- Publication Date
- 2026-07-24
AI Technical Summary
Existing technologies for setting parameters in orthokeratology lenses suffer from problems such as high lens misalignment and fitting failure rates due to spatial dimension compression loss and lack of high-dimensional feature resolution.
A multimodal chain labeling approach is adopted, which obtains a parameter rule base through convolutional neural network training, combines corneal topography and structured clinical parameters for dual-channel fusion, uses a CatBoost regression model for chain processing, and combines static image and dynamic video evaluation to adjust lens parameters to improve prediction accuracy.
It improves the accuracy and reliability of orthokeratology lens parameter prediction, reduces trial fitting costs, ensures the comprehensiveness and objectivity of evaluation results, and avoids prediction bias caused by single data.
Smart Images

Figure CN122451965A_ABST