| 周昊鹏,唐宇瑶,王常玺,李康,李真林.距离自注意力与长短期记忆模型预测CT设备X线管打火[J].中国医学影像技术,2025,41(4):659~665 |
| 距离自注意力与长短期记忆模型预测CT设备X线管打火 |
| Distance self-attention and long short-term memory model for predicting X-ray tube arcing in CT equipment |
| 投稿时间：2025-01-10 修订日期：2025-04-02 |
| DOI：10.13929/j.issn.1003-3289.2025.04.031 |
| 中文关键词: 深度学习 体层摄影术,X线计算机 自注意力 |
| 英文关键词:deep learning tomography, X-ray computed self-attention |
| 基金项目:四川大学华西医院"1·3·5工程"人工智能项目(ZYAI24031)、国家自然科学基金(12201441)。 |
|
| 摘要点击次数: 1062 |
| 全文下载次数: 109 |
| 中文摘要: |
| 目的 构建距离自注意力(DSA)与长短期记忆(LSTM)模型,观察其预测CT设备X线管打火的价值。方法 收集医疗物联网系统中的CT设备状态数据并进行预处理,构建基于模型注意(MA)模块和非线性衰减距离因子的DSA-LSTM模型,与其他模型比较,观察其预测CT设备X线管打火的价值。结果 相比其他模型,DSA-LSTM模型预测CT设备X线管打火的综合效能更佳。MA模块和非线性衰减距离因子可提高DSA-LSTM模型预测效能,且所有特征均对模型效能具有一定贡献。结论 DSA-LSTM模型可有效预测CT设备X线管打火。 |
| 英文摘要: |
| Objective To construct a distance self-attention (DSA) and long short-term memory (LSTM) model and observe its value for predicting X-ray tube arcing in CT equipment. Methods CT equipment status data of internet of medical things were collected and preprocessed, then DSA-LSTM model based on model attention (MA) module and nonlinear attenuation distance factor was constructed, and its value for predicting X-ray tube arcing in CT equipment was analyzed compared with other models. Results Compared with other models, DSA-LSTM model had better comprehensive efficiency for predicting X-ray tube arcing in CT equipment. MA module and nonlinear attenuation distance factor could improve the predictive efficiency of DSA-LSTM model, and all included features contributed to the performance of DSA-LSTM model in a certain extent. Conclusion DSA-LSTM model could effectively predict X-ray tube arcing in CT equipment. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
