A classification model generation method and device for predicting gestational hypertension
By using a two-stage classification model and deep learning technology, gestational hypertension can be identified using fundus images, which solves the problem of insufficient identification accuracy in existing technologies and achieves low-cost and efficient screening for gestational hypertension.
CN117152515BActive Publication Date: 2026-05-29BEIJING AIRDOC TECH CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- BEIJING AIRDOC TECH CO LTD
- Filing Date
- 2023-08-31
- Publication Date
- 2026-05-29
Smart Images
Figure CN117152515B_ABST
Abstract
The application discloses a kind of classification model generation method and device for predicting gestational hypertension, and the method one specific embodiment includes: carrying label fundus image as first training sample, obtain first sample dataset;Wherein, label includes normal label and gestational hypertension label;Gestational hypertension label includes: early gestational hypertension label, mild preeclampsia label and severe preeclampsia label;Based on first sample dataset, model training is generated two-classification model;And according to the gestational hypertension classification result obtained by second classification model to first sample dataset prediction, obtain second sample dataset;Then based on second sample dataset, model training is generated three-classification model;Finally, based on two-classification model and three-classification model, gestational hypertension classification model is generated.Therefore, the difficulty of model to the recognition of gestational hypertension is effectively reduced, and the accuracy of gestational hypertension classification is improved.
Need to check novelty before this filing date? Find Prior Art