Data-driven Method for 3D Axis-symmetric Object Reconstruction from Single Cone-beam Projection Data

Shousheng Luo, Ruyue Meng, Suhua Wei, Jian-Feng Cai, Xue-Cheng Tai, Yang Wang

Abstract

In this paper, we consider 3D axis-symmetric (AS) object reconstruction from single cone-beam x-ray projection data. Traditional x-ray CT fails to capture fleeting states of material due to the long time required for data acquisition at all angles. Therefore, AS objects are devised to investigate the instant deformation of material under pulse changes in the environment because single projection data is enough to reconstruct its inner structure. Previous reconstruction methods are layer-by-layer and ignore the longitudinal tilt of x-ray paths. We propose a regularization method using an adaptive tight frame to reconstruct the 3D AS object structure simultaneously. An alternating direction method is adopted to solve the proposed model. More importantly, a numerical algorithm is developed to compute the imaging matrix. Experiments on simulation data verify the effectiveness of our method.

Introduction

Traditional X-ray Computed Tomography (CT) requires rotating a source and detector around an object to capture hundreds of projections from different angles. This process is far too slow to capture "fleeting states"—instantaneous changes in materials caused by powerful shocks or explosions. To study these high-speed events, researchers often use axis-symmetric (AS) objects. Because these objects are symmetrical around a central axis, a single X-ray view theoretically contains enough information to reconstruct their entire 3D internal structure.

However, existing reconstruction methods typically process data "slice-by-slice." This approach ignores the longitudinal tilt of X-ray paths in a cone-beam system, leading to significant errors and artifacts, especially for longer objects. By treating a cone-beam as a series of 2D parallel slices, older methods fail to account for the true 3D geometry of the imaging system.

Methodology

The authors propose a simultaneous 3D reconstruction model that moves beyond the limitations of slice-by-slice processing. Their approach features three key components:

  1. Adaptive Wavelet Tight Frame (ATF): Instead of using a fixed mathematical basis to represent the image, the method "learns" an optimal filter bank directly from the data during the reconstruction process. This allows the model to better preserve sharp edges and fine details specific to the object being scanned.
  2. 3D Imaging Matrix Computation: A major technical contribution is an efficient numerical algorithm to calculate the "imaging matrix." This matrix represents the intersection lengths of every X-ray path with the 3D annular (ring-shaped) cylinders used to discretize the object's volume.
  3. Alternating Direction Method (ADM): To solve the complex optimization problem, the authors use ADM to break it into manageable sub-problems, iteratively updating the learned filters and the final attenuation map.
The cone-beam geometry used for axis-symmetric object reconstruction, showing the source S and the 3D annular discretization.

Experimental Results

The study relies on simulated data featuring a cylinder phantom with complex internal layers. To test robustness, Gaussian noise (mean 0, variance 0.03) was added to the projection data. The researchers compared their Adaptive Tight Frame (ATF) method against the standard Total Variation (TV) regularization method.

Quantitative comparison of Root Mean Square Error (RMSE) between TV and ATF methods across different lambda values.

As shown in the results table, the ATF method consistently outperformed TV across different regularization parameters (λ). For example, at λ = 0.005, the ATF method achieved an RMSE of 0.0278, compared to 0.0323 for the TV method. Visual inspections of the reconstructed cross-sections showed that the ATF method suppressed artifacts much more effectively, particularly in regions far from the central horizontal slice where X-ray tilt is most extreme.

Conclusion

By explicitly calculating the 3D path of tilted rays through annular elements and utilizing a self-learning regularization framework, this method eliminates the geometric distortions inherent in older slice-by-slice techniques. While the current iterative nature of the algorithm can be slow, it provides a highly flexible and accurate solution for capturing the internal structure of materials in fleeting states from a single projection.

Summary updated

Abstract

In this paper we consider 3D axis-symmetric (AS) object reconstruction from single cone-beam x-ray projection data. Traditional x-ray CT fails to capture fleeting state of material due to the long time for data acquisition at all angles. Therefore, AS object is devised to investigate the instant deformation of material under pulse change of environment because single projection data is enough to reconstruct its inner structure. Previous reconstruction methods are layer by layer, and ignore the longitudinal tilt of x-ray paths. We propose a regularization method using adaptive tight frame to reconstruct the 3D AS object structure simultaneously. Alternating direction method is adopted to solve the proposed model. More importantly, a numerical algorithm is developed to compute imaging matrix. Experiments on simulation data verify the effectiveness of our method
BibTeX
@inproceedings{luo2019data,
  author = {Shousheng Luo and Ruyue Meng and Suhua Wei and Jian-Feng Cai and Xue-Cheng Tai and Yang Wang},
  title = {{Data-driven Method for 3D Axis-symmetric Object Reconstruction from Single Cone-beam Projection Data}},
  booktitle = {Isicdm},
  year = {2019},
  doi = {10.1145/3364836.3364895},
}
Figure