Computed tomography super-resolution using convolutional neural networks

Haichao Yu, Ding Liu, Honghui Shi, Hanchao Yu, Zhangyang Wang, Xinchao Wang, Brent Cross, Matthew Bramler, Thomas S. Huang

Abstract

The practical application of Computed Tomography (CT) faces the dilemma between higher image resolution and less X-ray exposure for patients, motivating the research on CT super-resolution (SR). In this paper, we apply state-of-the-art SR techniques to reconstruct CT images using two proposed advanced CT SR models based on Convolutional Neural Networks (CNNs) and residual learning: a single-slice CT SR network (S-CTSRN), and a multi-slice CT SR network (M-CTSRN). S-CTSRN improves the high-frequency feature extraction by incorporating the residual learning strategy, while M-CTSRN further utilizes the coherence between neighboring CT slices for better SR reconstruction. We evaluate both models on a large-scale CT dataset <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup> , and obtain competitive results both quantitatively and qualitatively.
BibTeX
@inproceedings{yu2017computed,
  author = {Haichao Yu and Ding Liu and Honghui Shi and Hanchao Yu and Zhangyang Wang and Xinchao Wang and Brent Cross and Matthew Bramler and Thomas S. Huang},
  title = {{Computed tomography super-resolution using convolutional neural networks}},
  booktitle = {ICIP},
  year = {2017},
  doi = {10.1109/icip.2017.8297022},
}

Loading related content...

Figure