用单步滤波加速X射线散射张量重建，实现秒级实时成像。
Fast reconstruction of tensor tomographic X-ray scattering data for real-time applications
- 分离投影算子与视图相关混合算子，构建可推广的直接重建方法。
- 53×53×53×28张量体积重建仅需1秒，计算速度提升超10倍。
- 适用于多种散射模式，适合高通量或实时成像场景。
X射线散射张量断层扫描可揭示纳米尺度三维结构取向，但依赖慢速迭代重建限制了实时应用。本文提出一种平行束几何下直接重建方法的扩展，通过构造代数滤波器，在一次滤波与反投影中近似多个迭代更新。通过显式分离断层投影算子与视图相关混合算子，该方法可泛化至不同张量表示及小角X射线散射测量模式。模拟与实验结果表明，该方法在保持迭代重建精度的同时，计算时间减少一个数量级以上，可在商用硬件上实现53×53×53×28张量体在1秒内完成重建。显著的速度与稳定性使高通量分析成为可能，并为基于散射的实时成像开辟新路径。
原文摘要 · Abstract (English)
X-ray scattering tensor tomography reveals nanoscale structural orientation in 3D, but its reliance on slow iterative reconstruction limits real-time use. We introduce an extension of a direct reconstruction approach for parallel-beam geometries that computes algebraic filters approximating multiple iterative updates with a single filtering and back-projection step. By explicitly separating the tomographic projector from a view-dependent mixing operator, the method generalizes across tensor representations and modes of (small-angle) X-ray scattering measurement. Simulated and experimental results show that the approach approximates iterative reconstructions while reducing computation time by over an order of magnitude, realizing a reconstruction of a 53x53x53x28 tensor volume in 1 second on commercially-available hardware. The resulting speed and stability enable high-throughput analysis and open the door to real-time scattering-based imaging.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。