Quality Improvement of CT Scanning Using CNN-based Denoising of X-ray Transmission Images

DOI

Bibliographic Information

Other Title
  • CNNを用いたX線投影画像のノイズ除去に基づいたCTスキャンニングの品質向上

Description

<p>High precision 3D models can be used for various purpose like reverse engineering, flaw detection etc. However, to obatin such models, a large amount of projection images with low noise level are required and to get images with low noise level, the scanning time is very long. The purpose of my research is to denoise noisy CT images obtained from fast scanning, thus making them be eligible to be used for high precision reconstuction, reducing the total processing time. I applied deep learning approach to solve this problem, which I used a CNN with deep structure and two unique techniques: residual learning and batch normalization. The noise level of images are significantly reduced, and gained a more satisfying result than traditional methods.The total processing time is also reduced significantly.</p>

Journal

Keywords

Details 詳細情報について

  • CRID
    1390004222624863232
  • NII Article ID
    130007896528
  • DOI
    10.11522/pscjspe.2020s.0_202
  • Text Lang
    en
  • Data Source
    • JaLC
    • CiNii Articles
  • Abstract License Flag
    Disallowed

Report a problem

Back to top