XLXingyu Liu

Xingyu Liu

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Xingyu Liu has a total of 23 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Multi-view Diffusion Model with Gaussian Splatting in Projection Domain for Limited Angle Reconstruction

    by Xingyu Liu, Guangpu Yang, Steffen Kieß, Ct-Lab Ug, Jajnabalkya Guhathakurta, Sven Simon

    Industrial computed tomography (CT) plays a critical role in non-destructive testing and quality control; however, limited-angle acquisition often leads to severe artifacts due to restricted scanning rotation. To address this issue, we propose a novel framework that exploits the generative capability of a video diffusion model for multi-view generation in the projection domain and uses a novel reconstruction method. Specifically, a well-trained video diffusion model uses the available projections from limited-angle tomography as conditions to generate the missing-angle projections. To further enhance reconstruction quality and reduce the information loss introduced by traditional methods such as the Feldkamp–Davis–Kress (FDK) algorithm, we integrate a Gaussian splatting–based 3D reconstruction method, enabling volumetric reconstruction from the generated projections. Based on our experimental results, the proposed framework can generate novel projections of size 256 x 256 pixel in approximately 1 minute and complete the subsequent 3D reconstruction in about 5 minutes, offering new opportunities for inline CT applications. Experimental results demonstrate that novel-view generation substantially improves reconstruction performance, achieving significant gains in MSE, PSNR, and SSIM compared with conventional limited-angle tomography.


  • Open Access

    Compensating Streak Artifacts in Sparse-View Inline Industrial CT for Accurate Metrology using Self-Supervised Optimization of Implicit Neural Volume Representations

    by Faizan Ahmad, Guangpu Yang, Manuel Buchfink, Ammar Alsaffar, Ahmed Baraka, Xingyu Liu, Sven Simon

    Sparse-view computed tomography (CT) can reduce acquisition times, supporting inline industrial inspection in suitable settings. In practice, scan time may also be shortened by lowering exposure per view, using faster detectors or motion systems, or leveraging partial/parallel acquisition; here we focus on reducing the number of projections. Fewer projections, however, can introduce streak artifacts that cause measurement deviations during metrological evaluations. This paper presents two self-supervised deep learning approaches using implicit neural representations (INR) to mitigate sparse-view artifacts and enhance measurement accuracy. Both methods represent the 3D object volume using a multi-layer perceptron (MLP) optimized individually for each scan through an incremental forward-backward strategy. The first approach, Neural Representation with Sparse-View Volume-based Loss (NR-SVOL), employs volume-domain training using an initial filtered back-projection (FBP) volume, enabling rapid artifact reduction with limited computational overhead. The second, Neural Representation with Sparse-View Projection-based Loss (NR-SPRO), directly optimizes the INR to match measured sparse projections, analogous to Neural Radiance Fields (NeRF), yielding superior artifact compensation at the expense of increased computation. Comprehensive evaluations were conducted on three industrial objects, a gear, a cylinder head, and a connector, at varying sparse-view configurations (32–256 projections). Both NR-SVOL and NR-SPRO demonstrated substantial artifact reduction, decreasing surface deviations by up to an order of magnitude in standard deviation. NR-SVOL achieved results within approximately five minutes, suggesting compatibility with some inline cycle times for our tested parts, while NR-SPRO delivered even higher accuracy when allowed more computation. This study highlights a practical trade-off between speed and precision, showcasing the potential of these methods for sparse-view inline industrial CT for improved metrological quality.


  • Open Access

    A Fast Deep Incremental Angle Refinement Model for Limited-Angle CT Reconstruction

    by University of Stuttgart, Xingyu Liu, Tongji University, G.Q. Yang, Guangpu Yang, Ammar Alsaffar, Faizan Ahmad, Steffen Kieß, Sven Simon

    Industrial computed tomography (CT) plays a critical role in non-destructive testing and quality control across various industries. However, one of the key challenges in industrial CT lies in artifacts caused by limited-angle tomography, where geometric constraints prevent full rotation of the object. To mitigate these artifacts, we propose an innovative framework leveraging the diffusion model, a state-of-the-art generative model that has garnered significant attention for its powerful generation capabilities. Despite their impressive results, traditional diffusion models typically require upwards of 1,000 intermediate steps to achieve accurate output, leading to substantial computational demands and long inference time, which significantly limit their practical application. Our approach addresses these issues by integrating reconstructed data from different limited angles scans as intermediate steps, substituting the conventional steps of adding random Gaussian noise. This modification significantly reduces the number of intermediate steps necessary for both training and inference, thereby enhancing computational efficiency. Unlike other acceleration techniques that often compromise output quality for speed, our method achieves a significant reduction in computational burden while improving key performance metrics, such as structural similarity index (SSIM) and peak signal-to-noise ratio (PSNR). By aligning the training process more closely with real-world data degradation, our approach not only accelerates diffusion models but also enhances the fidelity and quality of the resulting reconstructions, making it a highly promising solution for industrial CT applications.


  • Open Access

    Compensating Sparse-view Inline Computed Tomography Artifacts with Neural Representation and Incremental Forward-Backward Network Architecture

    by University of Stuttgart, Manuel Buchfink, Faizan Ahmad, Guangpu Yang, Ammar Alsaffar, Charles Clark, RWTH Aachen University, Ahmed Baraka, Xingyu Liu, Sven Simon

    Sparse-view computed tomography (CT) is appealing for inline industrial quality control because it can significantly reduce scan acquisition times compared to full-view CT of 1-2 hours to a few minutes. This allows sparse view CT to be used in inline process manufacturing for 100% inspection of every component. However, the smaller number of projections in sparse-view CT scans leads to streak artifacts, which generally cause measurement deviations during the metrological inspection of components. Therefore, this paper discusses two deep-learning-based approaches for removing such artifacts. The two methods utilize a neural representation architecture to reproduce an estimate of the scanned object and then simulate a sparse-view forward projection to generate an artifact image. This approach was initially proposed for application in medical CT imaging, but it is modified for applications in inline industrial CT concerning speed. The two methods differ in how the neural representation is optimized during training. The first method is optimized based on an image-to-image basis, and the second method is on a projection basis. Our results show that both explored methods produce good correction results that lead to a promising minimization of surface deviation on the tested objects while meeting time constraints. This makes their application in industrial inline CT feasible.


  • Open Access

    Deep Incremental Angle Refinement Model for Limited-Angle CT Reconstruction - A Case Study on Concrete Specimens

    by University of Stuttgart, Xingyu Liu, Guangpu Yang, Ammar Alsaffar, Faizan Ahmad, Steffen Kieß, Sven Simon

    Industrial computed tomography (CT) is widely utilized for non-destructive testing and quality control in various industries. However, a common challenge in industrial CT is the presence of artifacts caused by limited angle tomography, where the object cannot be rotated fully due to geometric constraints or time limitations. To eliminate the artifacts, we propose a novel framework based on diffusion model: Deep Incremental Angle Refinement Model (DI-ARM). Our approach leverages the characteristics of CT projection by using reconstructed data of different limited angles as intermediate steps in the training process, replacing the traditional diffusion model of adding random Gaussian noise. This approach ensures data consistency in training process, mitigating the instability caused by sampling randomness of diffusion models. Furthermore, our method requires fewer steps compared to conventional diffusion models, significantly reducing computational resource consumption.


  • Open Access

    Artifact-robust Object Segmentation Using Thresholding Based on Binarized Image Object Analysis (TB2IOA) in X-ray Computed Tomography

    by University of Stuttgart, Xingyu Liu, Charles Clark, Steffen Kieß, Ammar Alsaffar, Hieu Tran, Guhathakurta Jajnabalkya, Sven Simon

    This paper presents a novel approach for image segmentation (e.g. pore segmentation) in computed tomography (CT) which is robust against typical artifacts in industrial CT like beam hardening or metal artifacts in general. In order to identify suited thre....