Scholarlink

This study proposes a web application to address the challenge of accessibility to Computed Tomography (CT) equipment by reconstructing 3D spatial information from a 2D X-ray image. Our proposed method is a multi-stage deep learning pipeline including generating synthetic data from DiffDRR, reducing the domain gap between synthetic and real images by CycleGAN, and reconstructing the CT image from X-ray2CTPA architecture. The results present that the proposed framework not only reconstructs 3D volumes with consistent anatomical structures but also allows users to interact with rotation and viewing slices (coronal, sagittal, axial) via the web interface. This study hopes to contribute to the collective effort of applying AI to enhance the informational value of common diagnostic imaging modalities, thereby opening up the potential for improved healthcare accessibility for the community.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Deep Learning-Based Web Application for CT Reconstruction from X-Ray Image

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Huu Quang Hoa Nguyen,
  • Tran Quoc Bao Le,
  • Quy Dat Tran,
  • Thi Tan Tien Nguyen,
  • Hang Phuong Nguyen,
  • Cong Tran,
  • Cuong Pham

摘要

This study proposes a web application to address the challenge of accessibility to Computed Tomography (CT) equipment by reconstructing 3D spatial information from a 2D X-ray image. Our proposed method is a multi-stage deep learning pipeline including generating synthetic data from DiffDRR, reducing the domain gap between synthetic and real images by CycleGAN, and reconstructing the CT image from X-ray2CTPA architecture. The results present that the proposed framework not only reconstructs 3D volumes with consistent anatomical structures but also allows users to interact with rotation and viewing slices (coronal, sagittal, axial) via the web interface. This study hopes to contribute to the collective effort of applying AI to enhance the informational value of common diagnostic imaging modalities, thereby opening up the potential for improved healthcare accessibility for the community.