ABSTRACT
Few studies present how to implement super resolution (SR) on abdominal Computed Tomography (CT) images. However, in the field of medical imaging, the development of an advanced SR system will potentially improve the clinical diagnosis and prognosis of various diseases, enabling the medical image generation machine to communicate its information better. The efficient sub-pixel convolutional network (ESPCN) is considered as the baseline network. Successfully implementing a residual learning strategy and proposing an approach for incorporating categorical conditions as the prior in the baseline, we proposed three novel deep residual learning based networks to restore more realistic texture for pancreatic tumour regions in abdominal CT images. The quality of super-resolved images is evaluated quantitatively by using Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index Measure (SSIM). Our proposed networks show better SR performance compared with other existing deep learning based methods on the abdominal CT dataset.
- Rachel E Jennings, Andrew A Berry, James P Strutt, David T Gerrard, and Neil A Hanley. Human pancreas development. Development, 142(18):3126–3137, 2015Google ScholarCross Ref
- S Prabu, V Balamurugan, and K Vengatesan. Design of cognitive image filters for suppression of noise level in medical images. Measurement, 141:296–301, 2019.Google ScholarCross Ref
- David A Bluemke, John L Cameron, Ralph H Hruban, Henry A Pitt, Stanley S Siegelman, Philippe Soyer, and Elliot K Fishman. Po- tentially resectable pancreatic adenocarcinoma: spiral ct assessment with surgical and pathologic correlation. Radiology, 197(2):381–385, 1995.Google ScholarCross Ref
- Chenyu You, Guang Li, Yi Zhang, Xiaoliu Zhang, Hongming Shan, Mengzhou Li, Shenghong Ju, Zhen Zhao, Zhuiyang Zhang, Wenxiang Cong, Ct super-resolution gan constrained by the identical, residual, and cycle learning ensemble (gan-circle). IEEE transactions on medical imaging, 39(1):188–203, 2019Google Scholar
- Wenzhe Shi, Jose Caballero, Ferenc Huszár, Johannes Totz, Andrew P Aitken, Rob Bishop, Daniel Rueckert, and Zehan Wang. Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 1874–1883, 2016.Google ScholarCross Ref
- Xintao Wang, Ke Yu, Chao Dong, and Chen Change Loy. Recovering realistic texture in image super-resolution by deep spatial feature transform. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 606–615, 2018.Google ScholarCross Ref
- Wenming Yang, Xuechen Zhang, Yapeng Tian, Wei Wang, Jing-Hao Xue, and Qingmin Liao. Deep learning for single image super-resolution: A brief review. IEEE Transactions on Multimedia, 21(12):3106–3121, 2019.Google ScholarDigital Library
- Chi-Hieu Pham. Deep learning for medical image super resolution and segmentation. PhD thesis, 2018.Google Scholar
- Chao Dong, Chen Change Loy, Kaiming He, and Xiaoou Tang. Image super-resolution using deep convolutional networks. IEEE transactions on pattern analysis and machine intelligence, 38(2):295–307, 2015.Google Scholar
- Guido F Montufar, Razvan Pascanu, Kyunghyun Cho, and Yoshua Bengio. On the number of linear regions of deep neural networks. In Advances in neural information processing systems, pages 2924–2932, 2014.Google Scholar
- Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016.Google ScholarCross Ref
- Christian Ledig, Lucas Theis, Ferenc Huszár, Jose Caballero, Andrew Cunningham, Alejandro Acosta, Andrew Aitken, Alykhan Tejani, Johannes Totz, Zehan Wang, Photo-realistic single image super-resolution using a generative adversarial network. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 4681–4690, 2017.Google ScholarCross Ref
- Xintao Wang, Ke Yu, Shixiang Wu, Jinjin Gu, Yihao Liu, Chao Dong, Yu Qiao, and Chen Change Loy. Esrgan: Enhanced super-resolution generative adversarial networks. In Proceedings of the European Conference on Computer Vision (ECCV), pages 0–0, 2018.Google Scholar
- Rushi Lan, Long Sun, Zhenbing Liu, Huimin Lu, Cheng Pang, and Xiaonan Luo. Madnet: A fast and lightweight network for single-image super resolution. IEEE transactions on cybernetics, 51(3):1443–1453, 2020.Google Scholar
- Marc A Attiyeh, Jayasree Chakraborty, Alexandre Doussot, Liana Langdon-Embry, Shiana Mainarich, Mithat Gönen, Vinod P Balachandran, Michael I D'Angelica, Ronald P DeMatteo, William R Jarnagin, Survival prediction in pancreatic ductal adenocarcinoma by quantitative computed tomography image analysis. Annals of surgical oncology, 25(4):1034–1042, 2018.Google Scholar
- GODFREY N Hounsfield. Computed medical imaging. nobel lecture, decemberr 8, 1979. Journal of computer assisted tomography, 4(5):665, 1980.Google Scholar
Index Terms
- Novel Approaches of Abdominal Computed Tomography Image Super Resolution based on Deep Residual Learning
Recommendations
Liver segmentation from computed tomography images using cascade deep learning
Abstract BackgroundLiver segmentation is a fundamental step in the treatment planning and diagnosis of liver cancer. However, manual segmentation of liver is time-consuming because of the large slice quantity and subjectiveness ...
Highlights- A computational method for liver segmentation in CT images is presented.
- Our ...
Registration of Planar Film Radiographs with Computed Tomography
MMBIA '96: Proceedings of the 1996 Workshop on Mathematical Methods in Biomedical Image Analysis (MMBIA '96)Abstract: The authors describe a method to register computed tomography (CT) data with planar film radiographs. Previous methods applied to the problem of CT-radiograph registration rely on determining the correspondence between occluding contours of ...
Brain atlas fusion from high-thickness diagnostic magnetic resonance images by learning-based super-resolution
AbstractIt is fundamentally important to fuse the brain atlas from magnetic resonance (MR) images for many imaging-based studies. Most existing works focus on fusing the atlases from high-quality MR images. However, for low-quality diagnostic images (...
Highlights- We fuse the brain atlas from real diagnostic MR images with high inter-slice thickness.
- All images are processed through the two-stage learning-based super-resolution.
- Groupwise registration is applied for unbiased atlas fusion.
Comments