一种放射学报告生成方法、装置、终端及存储介质
By combining a dynamic prior network model with a dynamic knowledge graph and a prior knowledge network, the problem of inconsistency between visual and textual data in radiology report generation is solved, enabling accurate description of rare abnormal areas and improving the quality of report generation.
CN117747042BActive Publication Date: 2026-07-17SHENZHEN INST OF ADVANCED TECH CHINESE ACAD OF SCI
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHENZHEN INST OF ADVANCED TECH CHINESE ACAD OF SCI
- Filing Date
- 2023-12-26
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
Existing radiology report generation technologies cannot fully cover and accurately describe rare abnormal areas, leading to inconsistencies between visual and textual data and affecting the quality of report generation.
Method used
A dynamic prior network model is adopted, which combines dynamic knowledge graph network and prior knowledge network to generate radiology reports through dynamic knowledge graph and prior knowledge information, thereby reducing text data bias and improving the quality of report generation.
Benefits of technology
By combining dynamic knowledge graphs and prior knowledge networks, the model can better handle visual and textual biases caused by limited data availability, thereby improving the accuracy and quality of radiology report generation.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN117747042B_ABST
Abstract
本发明所提供的一种放射学报告生成方法、装置、终端及存储介质，方法包括：将待处理的放射学图像输入预先训练的动态先验网络模型中，动态先验网络模型包括动态知识图谱网络、先验知识网络和解码器；在动态知识图谱网络中，根据放射学图像得到动态知识图谱，根据动态知识图谱得到视觉表征；获取放射学图像对应的先验知识信息，基于先验知识信息和视觉表征，利用先验知识网络得到先验表示向量；将视觉表征和先验表示向量输入解码器中，生成与放射学图像相对应的放射学报告。本发明通过结合动态知识图谱和先验知识信息生成放射学报告，能够更好地处理由于有限数据可用性而引起的视觉和文本偏见，提高了报告生成质量。
Need to check novelty before this filing date? Find Prior Art