A method of thyroid image classification and grading
By introducing a feature extraction network and a classifier, the problem of poor interpretability of the thyroid nodule identification model was solved. Multi-dimensional extraction and annotation of nodule features were achieved, generating interpretable evaluation results and improving the accuracy and interpretability of diagnosis.
CN122368561APending Publication Date: 2026-07-10RUIJIN HOSPITAL AFFILIATED TO SHANGHAI JIAO TONG UNIV SCHOOL OF MEDICINE
Patent Information
- Application Number
- CN202610176838.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-02-06
- Publication Date
- 2026-07-10
Smart Images
Figure CN122368561A_ABST
Abstract
This invention relates to the field of ultrasound image analysis technology, specifically to a method for classifying and rating thyroid images. The method includes: acquiring thyroid images from a patient; preprocessing the thyroid images to obtain preprocessed images; inputting the preprocessed images into a feature extraction network to obtain nodule extraction features; and feeding the nodule extraction features into corresponding classifiers to obtain corresponding evaluation results. Addressing the problem that existing thyroid nodule identification models are black-box and lack interpretability in assisted diagnosis, this invention introduces a specific feature extraction network to extract nodule features and echogenicity features across multiple dimensions. Each type of feature is labeled to facilitate subsequent diagnosis by doctors, and the output is based on relatively sufficient evidence and has strong interpretability.
Need to check novelty before this filing date? Find Prior Art
Citation Information
Patent Citations
Thyroid nodule classification method and system
CN119693673A
Thyroid nodule image classification method and imaging method
CN120526205A