一种基于蒙特卡罗和深度学习的质子成像散射校准方法
By modifying the physical processes and deep learning training of the Monte Carlo software, high-quality proton images were generated, solving the range aliasing problem and improving the accuracy of proton radiotherapy, especially for range monitoring before radiotherapy in lung cancer patients.
CN120459550BActive Publication Date: 2026-07-17HEFEI ION MEDICINE CENT
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HEFEI ION MEDICINE CENT
- Filing Date
- 2025-04-29
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
In existing proton imaging technology, the aliasing effect leads to a decrease in image quality, which cannot be effectively solved by existing scattering correction methods, thus affecting the accuracy of proton radiotherapy.
Method used
By modifying the physical processes in the Monte Carlo software, labeled images that can be used for training are generated, and end-to-end training is performed using deep learning methods to build a proton imaging scattering calibration model, thereby achieving scattering correction of proton images.
Benefits of technology
It improves the image quality of proton imaging, reduces the impact of range aliasing, and enhances the precision of proton radiotherapy, especially for range monitoring before radiotherapy in patients with lung tumors.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN120459550B_ABST
Abstract
本发明公开了一种基于蒙特卡罗和深度学习的质子成像散射校准方法，涉及医学影像、质子成像、深度学习和信号处理技术领域，方法为：在蒙特卡罗软件中对质子加速器进行建模；对蒙特卡罗软件进行拓展开发，实现质子角度偏转开 / 关的控制，从而实现质子在介质中有 / 无散射输运过程的控制；利用拓展开发后的蒙特卡罗软件构建能量解析剂量函数库，分别得到去散射前、后的能量解析剂量函数库；利用拓展开发后的蒙特卡罗软件和去散射前、后的能量解析剂量函数库进行质子成像，分别得到去散射前、后的质子图像；采集若干组质子成像的结果，构建样本集；利用样本集训练深度学习网络，输入为去散射前的质子图像，输出为去散射后的质子图像；训练完成后的深度学习网络用于对质子成像进行散射校准。
Need to check novelty before this filing date? Find Prior Art