Increase the resolution of the CT image to improve the detection accuracy of small tumours and other tumours, and thus enhance diagnostic precision. How to build a training dataset when paired high-resolution and low-resolution CT images are not available is what I will explore in this study. First of all, only low-resolution images are available; therefore, a degradation network and noise injection are employed to create domain-consistent low-resolution data. This way, the training data will be close to what you would find in real image pairs. Based on the above, I introduce a hierarchical attention transformer-based super-resolution network, DeHAT-SR. Shifted Window Self-Attention from Swin Transformer is employed to address the issue of long-range dependency. A network that adds channel attention and a shifted window cross-aggregation module can increase the receptive field and gather more pixel information for sharper super-resolution reconstruction. DeHAT-SR is a relatively stable adversarial training system that includes a generator, a discriminator and a feature extractor for super-resolution tasks. Thus, four CT images can be reconstructed at a high quality. The experiments have shown this. NIQE, BRISQUE and PIQE are no-reference image quality indicators; on 4x super-resolved CT images, DeHAT-SR outperforms the top methods such as EDSR, RCAN, ESRGAN and SwinIR. In addition, the visual comparison shows that DeHAT-SR produces images with finer details and is generally more perceptually pleasing.

Wenhui Xia, Huibo Wang, Tao Bian, Wenjiao Deng, Yunhe Li. Unsupervised Super-Resolution Analysis of CT Images Based on Hierarchical Attention Transformer. International Journal of World Medicine (2026), Vol. 7, Issue 1: 52-65. https://doi.org/10.38007/IJWM.2026.070106.

[1] T. M. Buzug, “Computed tomography,” in Springer handbook of medical technology: Springer, 2011, pp. 311-342.

[2] A. Hassan, S. A. Nazir, and H. Alkadhi, “Technical challenges of coronary CT angiography: today and tomorrow,” European journal of radiology, vol. 79, no. 2, pp. 161-171, 2011.

[3] H. M. Marshall, R. V. Bowman, I. A. Yang, K. M. Fong, and C. D. Berg, “Screening for lung cancer with low-dose computed tomography: a review of current status,” Journal of thoracic disease, vol. 5, no. Suppl 5, p. S524, 2013.

[4] J. Yang, J. Wright, T. Huang, and Y. Ma, “Image super-resolution as sparse representation of raw image patches,” in 2008 IEEE conference on computer vision and pattern recognition, 2008: IEEE, pp. 1-8.

[5] S. Gu, W. Zuo, Q. Xie, D. Meng, X. Feng, and L. Zhang, “Convolutional sparse coding for image super-resolution,” in Proceedings of the IEEE International Conference on Computer Vision, 2015, pp. 1823-1831.

[6] C. Osendorfer, H. Soyer, and P. Van Der Smagt, “Image super-resolution with fast approximate convolutional sparse coding,” in International Conference on Neural Information Processing, 2014: Springer, pp. 250-257.

[7] H. P. Babcock, J. R. Moffitt, Y. Cao, and X. Zhuang, “Fast compressed sensing analysis for super-resolution imaging using L1-homotopy,” Optics express, vol. 21, no. 23, pp. 28583-28596, 2013.

[8] C. Dong, C. C. Loy, K. He, and X. Tang, “Image super-resolution using deep convolutional networks,” IEEE transactions on pattern analysis and machine intelligence, vol. 38, no. 2, pp. 295-307, 2015.

[9] B. Lim, S. Son, H. Kim, S. Nah, and K. M. Lee, “Enhanced deep residual networks for single image super-resolution,” in Proceedings of the IEEE conference on computer vision and pattern recognition workshops, 2017, pp. 136-144.

[10] Y. Zhang, K. Li, K. Li, L. Wang, B. Zhong, and Y. Fu, “Image super-resolution using very deep residual channel attention networks,” in Proceedings of the European conference on computer vision (ECCV), 2018, pp. 286-301.

[11] I. J. Goodfellow et al., “Generative adversarial networks,” arXiv preprint arXiv:1406.2661, 2014.

[12] C. Ledig et al., “Photo-realistic single image super-resolution using a generative adversarial network,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 4681-4690.

[13] X. Wang et al., “Esrgan: Enhanced super-resolution generative adversarial networks,” in Proceedings of the European Conference on Computer Vision (ECCV) Workshops, 2018, pp. 0-0.

[14] X. Jiang, Y. Xu, P. Wei, and Z. Zhou, “CT Image Super Resolution Based On Improved SRGAN,” in 2020 5th International Conference on Computer and Communication Systems (ICCCS), 15-18 May 2020, pp. 363-367.

[15]  X. Zhang, C. Feng, A. Wang, L. Yang, and Y. Hao, “CT super-resolution using multiple dense residual block based GAN,” Signal, Image and Video Processing, vol. 15, no. 4, pp. 725-733, 2021.

[16] A. Dosovitskiy et al., “An image is worth 16x16 words: Transformers for image recognition at scale,” arXiv preprint arXiv:2010.11929, 2020.

[17] Z. Liu, Y. Lin, Y. Cao, H. Hu, Y. Wei, Z. Zhang, S. Lin, and B. Guo, “Swin transformer: Hierarchical vision transformer using shifted windows,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 10012-10022.

[18] J. Liang, J. Cao, G. Sun, K. Zhang, L. Van Gool, and R. Timofte, “SwinIR: Image restoration using Swin Transformer,” in Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops, 2021, pp. 1833-1844.

[19] X. Chen, X. Wang, J. Zhou, Y. Qiao, and C. Dong, “Activating more pixels in image super-resolution transformer,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 22367-22377.

[20] M. V. Conde, U.-J. Choi, M. Burchi, and R. Timofte, “Swin2SR: SwinV2 transformer for compressed image super-resolution and restoration,” in Proceedings of the European Conference on Computer Vision Workshops, 2022.

[21] Z. Chen, Y. Zhang, J. Gu, L. Kong, X. Yang, and F. Yu, “Dual aggregation transformer for image super-resolution,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 12202-12211.

[22] F. Wang, H. Li, R. Zhang, et al., “CTformer: Convolution-free Token2Token dilated vision transformer for low-dose CT denoising,” Physics in Medicine & Biology, vol. 68, 2023.

[23] Y. Chu, L. Zhou, G. Luo, et al., “HorusEye: A self-supervised foundation model for generalizable X-ray tomography restoration,” Nature Computational Science, 2026.

[24] Comparison of conventional 512-matrix CT images with Swin2SR-based 2048-matrix images in the visualization and diagnosis of lung nodules, Japanese Journal of Radiology, 2026.

[25] R. Keys, “Cubic convolution interpolation for digital image processing,” IEEE transactions on acoustics, speech, and signal processing, vol. 29, no. 6, pp. 1153-1160, 1981.

[26] K. Simonyan and A. Zisserman, “Very deep convolutional networks for large-scale image recognition,” arXiv preprint arXiv:1409.1556, 2014.

[27] U. Demir and G. Unal, “Patch-based image inpainting with generative adversarial networks,” arXiv preprint arXiv:1803.07422, 2018.

[28] K. Clark et al., “The Cancer Imaging Archive (TCIA): maintaining and operating a public information repository,” Journal of digital imaging, vol. 26, no. 6, pp. 1045-1057, 2013.

[29] D. Goldgof et al., Data From QIN_LUNG_CT, doi: https://doi.org/10.7937/K9/TCIA.2015.NPGZYZBZ.

[30] J. Kalpathy-Cramer, S. Napel, D. Goldgof, and B. Zhao, QIN multi-site collection of Lung CT data with Nodule Segmentations,

[31] A. Mittal, R. Soundararajan, and A. C. Bovik, “Making a ‘completely blind’ image quality analyzer,” IEEE Signal processing letters, vol. 20, no. 3, pp. 209-212, 2012.

[32] A. Mittal, A. K. Moorthy, and A. C. Bovik, “No-reference image quality assessment in the spatial domain,” IEEE Transactions on image processing, vol. 21, no. 12, pp. 4695-4708, 2012.

[33] N. Venkatanath, D. Praneesh, M. Lamb, B. J. Philip, and S. S. Iyengar, “Blind image quality evaluation using perception based local features,” in 2013 IEEE International Conference on Acoustics, Speech and Signal Processing, 2013, pp. 2207-2211.

[34] X. Wang et al., “BasicSR: Open source image and video restoration toolbox,” https://github.com/XPixelGroup/BasicSR, 2022.