Task-driven sparse-view planar X-ray computed tomography for axial metrology
Organizations: SKL-Radio Frequency Heterogeneous Integration, Shenzhen University, Shenzhen 518060, China · SKL-Precision Micro-Nano Manufacturing Technology, Xi’an Jiaotong University, Xi’an 710000, China · Visual Computing Center, King Abdullah University of Science and Technology, Thuwal 23955-6900, Saudi Arabia
Abstract
Computational imaging can reduce the acquisition burden when the system is designed around task-relevant information rather than complete image recovery. We investigate this framework for sparse-view planar X-ray tomography (XCT) of layered objects, using measurement of the axial termination gap (ATG) of back-drilled printed-circuit-board (PCB) vias as a representative task. Instead of using 3D reconstruction, the proposed method estimates ATG directly from polychromatic projection data through a physics-based sensitivity model. A gated hierarchical estimator first fits ATG together with radiometric perturbations and selectively introduces global geometry and specimen-warp models only when a reduced chi-squared adequacy test and Bayesian information criterion justify the added degrees of freedom. The same low-dimensional perturbation model is incorporated into marginalized Fisher-information analyses, yielding task-specific conditional and calibration-aware Cram'er-Rao bounds. Numerical experiments show that, under a fixed total photon budget, task-driven metrology changes little over 12-64 views, while reconstruction-driven metrology is strongly view-dependent with much larger errors. The direct estimator also follows the expected photon-limited scaling. These results demonstrate a computational sensing strategy in which XCT acquisition, physical modeling, and inference are evaluated according to the quantitative information required by the task rather than for general-purpose volumetric image fidelity.
Figures & tables
| Condition | Bias | SD | RMSE | RMSE CI | Pass rate | GA rate | WA rate | |
|---|---|---|---|---|---|---|---|---|
| Matched | -0.0175 | 0.1518 | 0.1503 | 0.1181–0.1791 | 1.0000 | 1.0000 | 0.0000 | 0.0000 |
| Spectrum | -0.0195 | 0.1118 | 0.1117 | 0.0920–0.1305 | 1.0003 | 1.0000 | 0.0000 | 0.0000 |
| Blur | -0.0271 | 0.1960 | 0.1946 | 0.1474–0.2386 | 1.0000 | 1.0000 | 0.0000 | 0.0000 |
| Scatter | 0.0028 | 0.1467 | 0.1443 | 0.1024–0.1843 | 0.9998 | 1.0000 | 0.0000 | 0.0000 |
| GEO global | 0.1034 | 0.1148 | 0.1530 | 0.1212–0.1819 | 1.0119 | 1.0000 | 1.0000 | 0.0000 |
| Board warp | 0.0239 | 0.1364 | 0.1362 | 0.1034–0.1679 | 0.9999 | 1.0000 | 0.1333 | 1.0000 |
| Cat | Parameter | Nominal value / range | Description |
| Object/task | Lateral ROI | Canonical layered phantom | |
| Object thickness | Canonical layered phantom | ||
| Axial gap (ATG), | Reference plane to feature termination | ||
| Reference plane, | Axial location of the Cu reference layer | ||
| Cu layer thickness | Thickness of the Cu reference layer | ||
| Material classes | Cu and FR4 | Two effective PCB material classes |
| Cat | Parameter | Nominal value / range | Description |
| Radiometric | Scatter fraction, | nominal; perturbation | Amplitude of the low-frequency additive scatter component |
| Effective blur, | px nominal; px perturbation | Effective combined detector/focal-spot blur | |
| Tube voltage, | kVp nominal; kVp perturbation | Effective spectral-endpoint perturbation | |
| Cu filtration, | mm nominal; mm perturbation | Effective filtration perturbation used to represent spectral perturbation. | |
| Geometric | Relative magnification, | nominal | Global magnification perturbation . |
| Global perturbation, | Effective relative magnification change |
Computed Tomography
Computed tomography uses X-ray measurements from multiple angles to reconstruct cross-sectional and volumetric images for diagnosis and scientific imaging.
See all 288 papers