R ^2 2 -Net: Recurrent and Recursive Network for Sparse-View CT Artifacts Removal

Tiancheng Shen, Xia Li, Zhisheng Zhong, Jianlong Wu, Zhouchen LinMICCAI 2019 · OpenReview

Abstract

We propose a novel neural network architecture to reduce streak artifacts generated in sparse-view 2D Computed Tomography image reconstruction. This architecture decomposes the streak artifacts removal into multiple stages through the recurrent mechanism, which can fully utilize information in previous stages and guide the learning of later stages. In each recurrent stage, the key components of the architecture operate recursively. The recursive mechanism is helpful to save parameters and enlarge the receptive field efficiently with exponentially increased dilation of convolution. To verify its effectiveness, we conduct experiments on the AAPM’s CT dataset through 5-fold cross-validation. Our proposed method outperforms the state-of-the-art methods both quantitatively and qualitatively.
BibTeX
@inproceedings{shen2019r,
  author = {Tiancheng Shen and Xia Li and Zhisheng Zhong and Jianlong Wu and Zhouchen Lin},
  title = {{R ^2 2 -Net: Recurrent and Recursive Network for Sparse-View CT Artifacts Removal}},
  booktitle = {International Conference on Medical Image Computing and Computer-Assisted Intervention},
  year = {2019},
  url = {https://openreview.net/forum?id=bt8h_0B7HU},
}

Loading related content...

Figure