ABSTRACT
Reducing low-dose radiation while maintaining high-quality image reconstruction in X-ray computed tomography (CT) is challenging, due to the reconstruction images degradation as the number of projection view decreases. As opposed to most of the existing deep network approaches with supervised learning scheme, which requires the data in the learning and testing procedure to be the same dimension, in the paper we propose an unsupervised learning approach based on the denoising autoencoding prior (DAEP) for few-view CT reconstruction. Two innovations were done to substantially improve the naive DAEP. First, by employing the virtual variables technique, higher-dimensional network is learned and then incorporated into the single-channel reconstruction procedure. Second, replacing the L2 regression loss function by more robust L1 regression is favor to preserve texture details. Experimental results demonstrate that the proposed method can achieve promising effects over state-of-the-art methods in terms of peak signal-to-noise ratio (PSNR) and subjective visual quality.
- A. J. Einstein, M. J. Henzlova, and S, Rajagopalan. 2007. Estimating risk of cancer associated with radiation exposure from 64-slice CT coronary angiography. J. Am. Med. Assoc. 298(3), 317--323.Google ScholarCross Ref
- D. J. Brenner and E. J. Hall. 2007. CT-An increasing source of radiation exposure. New Eng. J. Med. 357(22), 2277--2284.Google ScholarCross Ref
- Y. Liu, J. Ma, Y. Fan, and Z. Liang. 2012. Adaptive-weighted total variation minimization for sparse data toward low-dose X-ray computed tomography image reconstruction. Phys Med Biol. 57(23), 7923--7956.Google ScholarCross Ref
- R. Gordon, R. Bender, and G. T. Herman. 1970. Algebraic Reconstruction Techniques (ART) for three-dimensional electron microscopy and X-ray photography. Journal of Teoretical Biology. 29(3), 471--481.Google ScholarCross Ref
- A. H. Andersen and A. C. Kak. 1984. Simultaneous algebraic reconstruction technique (SART): a superior implementation of the art algorithm. Ultrasonic Imaging. 6(1), 81--94.Google ScholarCross Ref
- E. Y. Sidky, C. M. Kao, and X. Pan. 2006. Accurate image reconstruction from few-views and limited-angle data in divergent-beam CT. Journal of X-ray Science and Technology. 14(2), 119--139.Google Scholar
- J. Song, Q. H. Liu, G. A. Johnson, and C. T. Badea. 2007. Sparseness prior based iterative image reconstruction for retrospectively gated cardiac micro-CT. Medical physics. 34(11), 4476--4483.Google Scholar
- Z. Tian, X. Jia, K. Yuan, T. Pan, and S. B. Jiang. 2011. Low-dose CT reconstruction via edge-preserving total variation regularization. Phys Med Biol. 56(18), 5949--5967.Google ScholarCross Ref
- Y Liu, Z Liang, J Ma, H Lu, and K Wang. 2014. Total variation-stokes strategy for sparse-view X-ray CT image reconstruction. IEEE Trans. Med. Imag. 33(3), 749--763.Google ScholarCross Ref
- Q. Xu, H. Yu, X. Mou, and G. Wang. 2014. Dictionary learning based low dose x-ray CT reconstruction. Frontiers of Medical Imaging. 99--119.Google Scholar
- Z. Hu, Q. Liu, N. Zhang, Y. Zhang, X. Peng, P. Wu, and D. Liang. 2016. Image reconstruction from few-view CT data by gradient-domain dictionary learning. Journal of X-ray science and technology. 24(4), 627--638.Google ScholarCross Ref
- O. Ronneberger, P. Fischer, and T. Brox. 2015. U-Net: Convolutional networks for biomedical image segmentation. In International Conference on Medical Image Computing and Computer-Assisted Intervention. Springer. 234--241.Google Scholar
- J. Kim, J. Kwon Lee, and K. Mu Lee. 2016. Deeply-recursive convolutional network for image super-resolution. In Proc. IEEE CVPR. 1637--1645.Google Scholar
- T. Wurfl, F. C. Ghesu, V. Christlein, and A. Maier. 2016. Deep learning computed tomography. In International Conference on MICCAI. Springer. 432--440.Google Scholar
- E. Kang, J. Min, and J. C. Ye. 2017. A deep convolutional neural network using directional wavelets for low-dose x-ray CT reconstruction. Medical Physics. 44(10), 360--375.Google ScholarCross Ref
- H. Chen, Y. Zhang, W. Zhang, P. Liao, K. Li, J. Zhou, and G. Wang. 2017. Low-dose CT via convolutional neural network. Biomedical optics express. 8(2), 679--694.Google Scholar
- H. Chen, Y. Zhang, M. K. Kalra, F. Lin, P. Liao, J. Zhou, and G. Wang. 2017. Low-dose CT with a residual encoder-decoder convolutional neural network (RED-CNN). arXiv preprint arXiv:1702.00288.Google Scholar
- E. Kang and J. C. Ye. 2017. Wavelet domain residual network (WavResNet) for low-dose X-ray CT reconstruction. International Meeting on Fully Three-Dimensional Image Reconstruction in Radiology and Nuclear Medicine (arXiv preprint arXiv:1703.01383).Google Scholar
- E. Kang, J. Yoo, and J. C. Ye.2017. Wavelet residual network for low-dose CT via deep convolutional framelets. ArXiv preprint arXiv:1707.09938.Google Scholar
- K. H. Jin, M. T. McCann, E. Froustey, and M. Unser. 2017. Deep convolutional neural network for inverse problems in imaging. IEEE Trans. Image Process. 26(8), 4509--4522.Google ScholarCross Ref
- Y. Han, J. Yoo, and J. C. Ye. 2016. Deep residual learning for compressed sensing CT reconstruction via persistent homology analysis. ArXiv preprint arXiv:1611.06391.Google Scholar
- S.A. Bigdeli and M. Zwicker. 2017. Image restoration using autoencoding priors. ArXiv preprint arXiv: 1703.09964.Google Scholar
- G. Alain and Y. Bengio. 2014. What regularized auto-encoders learn from the data-generating distribution. Journal of Machine Learning Research. 15, 3743--3773. Google ScholarDigital Library
- N. Joshi, C. L. Zitnick, R. Szeliski, D. Kriegman. 2009. Image deblurring and denoising using color priors. Proc. IEEE Conf. Comput. Vis. Pattern Recognit. 1550--1557.Google ScholarCross Ref
- H. S. Mousavi and V. Monga. 2017. Sparsity-based color image super resolution via exploiting cross channel constraints, IEEE Trans. Image Process. 26(11), 5094--5106.Google ScholarDigital Library
- Q. Liu, H. Leung. 2018. Variable augmented neural network for decolorization and multi-exposure fusion. Information Fusion. 46, 114--127.Google ScholarCross Ref
- K.N. Chaudhury, and M. Unser. 2012. Non-local Euclidean medians. IEEE Signal Processing Letters. 19(11), 745--748.Google ScholarCross Ref
- I. Gorodnitsky, and B. Rao. 1997. Sparse signal reconstruction from limited data using FOCUSS: a re-weighted minimum norm algorithm. IEEE Trans. Signal Process. 45(3), 600--616. Google ScholarDigital Library
- B.D. Rao, and K.K-. Delgado.1999. An affine scaling methodology for best basis selection. IEEE Trans. Signal Process. 47(1), 187--200. Google ScholarCross Ref
- P. Rodriguez, B. Wohlberg. 2009. Efficient minimization method for a generalized total variation functional. IEEE Trans. Image Process. 18(2), 322--332. Google ScholarDigital Library
- P. Rodriguez, and B. Wohlberg. 2008. An efficient algorithm for sparse representations with lp data fidelity term. In: Proceedings of 4th IEEE Andean Technical Conference (ANDESCON).Google Scholar
- J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and F. Li. 2009. Imagenet: A large-scale hierarchical image database. IEEE Trans. Image Process. 248--255.Google Scholar
- J. Cong and B. Xiao. 2014. Minimizing computation in convolutional neural networks. In Proc. Int. Conf. Artif. Neural Netw. 281--290.Google Scholar
- Y. Han and J. Ye. June, 2018. Framing U-Net via deep convolutional framelets: Application to sparse-view CT. IEEE Trans. Medical Imag. 37(6), 1418--1429Google ScholarCross Ref
Index Terms
- Sparse-View CT Reconstruction via Robust and Multi-channels Autoencoding Priors
Recommendations
Learning Projection Views for Sparse-View CT Reconstruction
MM '22: Proceedings of the 30th ACM International Conference on MultimediaSparse-View CT (SVCT), which provides low-dose and high-speed CT imaging, plays an important role in the medical imaging area. As the decrease of projection views, the reconstructed image suffers from severe artifacts. To this end, recent works utilize ...
A Transformer-Based Iterative Reconstruction Model for Sparse-View CT Reconstruction
Medical Image Computing and Computer Assisted Intervention – MICCAI 2022AbstractSparse-view computed tomography (CT) is one of the primary means to reduce the radiation risk. But the reconstruction of sparse-view CT will be contaminated by severe artifacts. By carefully designing the regularization terms, the iterative ...
A dual-domain neural network based on sinogram synthesis for sparse-view CT reconstruction
Highlights- To suppress or remove the streak artifacts in parse-view computed tomography (CT) images, a convolutional neural network is employed to synthesize the full-...
Abstract ObjectiveThe dual-domain deep learning-based reconstruction techniques have enjoyed many successful applications in the field of medical image reconstruction. Applying the analytical reconstruction based operator to ...
Comments