Pelvic fracture probability prediction method, system, medium, program product and terminal based on deep learning
By segmenting and analyzing pelvic CT images using a deep learning-based pelvic segmentation network and fracture probability prediction model, the problems of low computational efficiency and low accuracy of deep learning models in medical image analysis are solved, achieving high efficiency and accuracy in fracture detection and supporting automated diagnosis and treatment of pelvic fractures.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- YINGWEI MEDICAL TECHNOLOGY (SUZHOU) CO LTD
- Filing Date
- 2024-12-10
- Publication Date
- 2026-06-19
AI Technical Summary
In existing medical image analysis tasks, deep learning models suffer from low computational efficiency and low analytical accuracy when processing large-scale medical image data. Furthermore, they exhibit poor generalization ability and robustness, leading to performance degradation and inaccurate diagnosis across different hospitals, equipment, or patient groups.
A deep learning-based pelvic segmentation network was used to segment pelvic CT images. The Triplet-Attention-Unet algorithm was used for accurate segmentation. A pelvic fracture probability prediction model was constructed by combining a 3D Vnet network. Fracture probability was predicted by bone density analysis and bone quality analysis, and a fracture probability prediction distribution map was generated.
It has improved the accuracy and efficiency of fracture detection and analysis, reduced misdiagnosis and missed diagnosis, achieved automated processing, improved processing speed and prediction accuracy, and provided a scientific basis for the diagnosis and treatment of pelvic fractures.
Smart Images

Figure CN122243851A_ABST