X射線冠狀動脈造影（XCA）是診斷心血管疾病的重要工具。然而，由於圖像解析度不足、視覺評估過於主觀且缺乏量化。因此，本文探討了超解析（SR）技術如何幫助提升 XCA 圖像的質量。我們發現，通過遷移學習技術和選擇當前適合的 SR 模型，可以顯著改善XCA圖像的質量。然而，當 SR 增強的圖像用於血管分割時，可能會產生不良輸出。為了解決這個問題，我們提出了一個新模型 super resolution enhanced TransUNet（SRE-TransUNet），該模型結合了 TransUNet 的架構和高解析度的 XCA 圖像。實驗結果表明，SRE-TransUNet 結合 Real-ESRGAN+ 在 CAG 數據集的模糊圖像上可以達到88.30%的F1分數，比現有技術方法高出1.02%。

X-ray coronary angiography (XCA) is an important tool to diagnose the cardiovascular diseases. However, due to insufficient image resolution, visual assessment is subjective and lacks quantification. In this paper, we investigate how the super-resolution (SR) technology can help the image enhancement for XCA. We found that with transfer learning techniques and carefully selection of SR models, the quality of XCA images can be improved significantly. However, when the SR enhanced images are utilized for vessel segmentation, they could produce undesired outputs. To solve this problem, we proposed a new model, called Super-Resolution Enhanced TransUNet (SRE-TransUNet), which combines the architecture of TransUNext and the high-resolution XCA images. Experimental results show that SRE-TransUNet with Real-ESRGAN+ can reach F1 score 88.30%, on blurry images of CAG dataset, which is 1.02% higher than state-of-the-art methods.

[1] Michela Di Cesare et al. “The Heart of the World”. Global Heart 19.1 (2024),
p. 11. doi: 10.5334/gh.1288.
[2] David J Piñeiro et al. “World Heart Day 2023: Knowing your heart”. Indian
J Med Res 158.3 (2023), pp. 213–215. doi: \url{10.4103/ijmr.ijmr_
1689_23}.
[3] Michael R. Rudnick. Prevention of contrast-associated acute kidney injury related to angiography. Ed. by Paul M. Palevsky. UpToDate. Updated Nov 27,
2023. 2023. url: https://www.uptodate.com/contents/preventionof-contrast-associated-acute-kidney-injury-related-toangiography (visited on 05/2024).
[4] Stella K. Kang. Diagnosis and treatment of an acute reaction to a radiologic
contrast agent. Ed. by Jr N. Franklin Adkinson and Anna M. Feldweg. UpToDate. Updated Jun 06, 2024. 2024. url: https://www.uptodate.com/
contents/diagnosis-and-treatment-of-an-acute-reactionto-a-radiologic-contrast-agent (visited on 06/26/2024).
[5] Matthew S. Davenport and Jeffrey H. Newhouse. Patient evaluation prior to
oral or iodinated intravenous contrast for computed tomography. Ed. by Herbert Y. Kressel. UpToDate. Updated May 14, 2024. 2024. url: https : / /
www.uptodate.com/contents/patient- evaluation- prior- tooral - or - iodinated - intravenous - contrast - for - computed -
tomography (visited on 06/26/2024).
[6] Hayit Greenspan. “Super-Resolution in Medical Imaging”. The Computer Journal 52.1 (Feb. 2008), pp. 43–63. issn: 0010-4620. doi: 10.1093/comjnl/
bxm075. eprint: https : / / academic . oup . com / comjnl / article -
pdf/52/1/43/1153004/bxm075.pdf. url: https://doi.org/10.
1093/comjnl/bxm075.
[7] Xintao Wang et al. “Real-ESRGAN: Training Real-World Blind Super-Resolution
with Pure Synthetic Data”. 2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW). 2021, pp. 1905–1914. doi: 10 . 1109 /
ICCVW54120.2021.00217.
[8] Jianyi Wang et al. Exploiting Diffusion Prior for Real-World Image SuperResolution. 2023. arXiv: 2305.07015 [cs.CV]. url: https://arxiv.
org/abs/2305.07015.
[9] Marcos V. Conde et al. “Swin2SR: SwinV2 Transformer for Compressed Image
Super-Resolution and Restoration”. Computer Vision – ECCV 2022 Workshops.
Ed. by Leonid Karlinsky, Tomer Michaeli, and Ko Nishino. Cham: Springer
Nature Switzerland, 2023, pp. 669–687. isbn: 978-3-031-25063-7.
[10] Zongsheng Yue, Jianyi Wang, and Chen Change Loy. “ResShift: efficient diffusion model for image super-resolution by residual shifting”. Proceedings of the
37th International Conference on Neural Information Processing Systems. NIPS
’23. New Orleans, LA, USA: Curran Associates Inc., 2024.
[11] Lingchen Sun et al. Improving the Stability of Diffusion Models for Content
Consistent Super-Resolution. 2023. arXiv: 2401 . 00877 [eess.IV]. url:
https://arxiv.org/abs/2401.00877.
[12] Junjie Ke et al. “MUSIQ: Multi-scale Image Quality Transformer”. 2021 IEEE/CVF
International Conference on Computer Vision (ICCV). 2021, pp. 5128–5137.
doi: 10.1109/ICCV48922.2021.00510.
[13] Sidi Yang et al. “MANIQA: Multi-dimension Attention Network for No-Reference
Image Quality Assessment”. 2022 IEEE/CVF Conference on Computer Vision
and Pattern Recognition Workshops (CVPRW). 2022, pp. 1190–1199. doi: 10.
1109/CVPRW56347.2022.00126.
[14] Jianyi Wang, Kelvin C.K. Chan, and Chen Change Loy. “Exploring CLIP for
assessing the look and feel of images”. Proceedings of the Thirty-Seventh AAAI
Conference on Artificial Intelligence and Thirty-Fifth Conference on Innovative
Applications of Artificial Intelligence and Thirteenth Symposium on Educational
Advances in Artificial Intelligence. AAAI’23/IAAI’23/EAAI’23. AAAI Press,
2023. isbn: 978-1-57735-880-0. doi: 10.1609/aaai.v37i2.25353. url:
https://doi.org/10.1609/aaai.v37i2.25353.
[15] Alejandro Frangi et al. “Multiscale Vessel Enhancement Filtering”. Med. Image
Comput. Comput. Assist. Interv. 1496 (Feb. 2000).
[16] Olaf Ronneberger, Philipp Fischer, and Thomas Brox. “U-Net: Convolutional
Networks for Biomedical Image Segmentation”. Medical Image Computing and
Computer-Assisted Intervention – MICCAI 2015. Ed. by Nassir Navab et al.
Cham: Springer International Publishing, 2015, pp. 234–241. isbn: 978-3-319-
24574-4.
[17] Hu Cao et al. “Swin-Unet: Unet-Like Pure Transformer for Medical Image Segmentation”. Computer Vision – ECCV 2022 Workshops. Ed. by Leonid Karlinsky, Tomer Michaeli, and Ko Nishino. Cham: Springer Nature Switzerland,
2023, pp. 205–218. isbn: 978-3-031-25066-8.
[18] Jieneng Chen et al. “TransUNet: Transformers Make Strong Encoders for Medical Image Segmentation”. ArXiv abs/2102.04306 (2021). url: https://api.
semanticscholar.org/CorpusID:231847326.
[19] R. Keys. “Cubic convolution interpolation for digital image processing”. IEEE
Transactions on Acoustics, Speech, and Signal Processing 29.6 (1981), pp. 1153–
1160. doi: 10.1109/TASSP.1981.1163711.
[20] Chao Dong et al. “Image Super-Resolution Using Deep Convolutional Networks”.
IEEE Transactions on Pattern Analysis and Machine Intelligence 38.2 (2016),
pp. 295–307. doi: 10.1109/TPAMI.2015.2439281.
[21] Yan Hu et al. “SuperVessel: Segmenting High-Resolution Vessel from LowResolution Retinal Image”. Pattern Recognition and Computer Vision. Ed. by
Shiqi Yu et al. Cham: Springer Nature Switzerland, 2022, pp. 178–190. isbn:
978-3-031-18910-4.
[22] Zhonghua Sun and Curtise K.C. Ng. “Finetuned Super-Resolution Generative
Adversarial Network (Artificial Intelligence) Model for Calcium Deblooming
in Coronary Computed Tomography Angiography”. Journal of Personalized
Medicine 12 (Aug. 2022), p. 1354. doi: 10.3390/jpm12091354.
[23] Makoto Orii et al. Super-resolution deep learning reconstruction at coronary
computed tomography angiography to evaluate the coronary arteries and instent lumen: An initial experience. July 2022. doi: 10.21203/rs.3.rs1875541/v2.
[24] Tim Jerman et al. “Enhancement of Vascular Structures in 3D and 2D Angiographic Images”. IEEE Transactions on Medical Imaging 35.9 (2016), pp. 2107–
2118. doi: 10.1109/TMI.2016.2550102.
[25] Shoujun Zhou et al. “Automatic segmentation of coronary angiograms based
on fuzzy inferring and probabilistic tracking”. Biomedical engineering online 9
(Aug. 2010), p. 40. doi: 10.1186/1475-925X-9-40.
[26] Rashindra Manniesing, Max Viergever, and W.J. Niessen. “Vessel Axis Tracking
Using Topology Constrained Surface Evolution”. Medical Imaging, IEEE Transactions on 26 (Apr. 2007), pp. 309 –316. doi: 10.1109/TMI.2006.891503.
[27] Jingyang Zhang et al. “Vesselness-constrained robust PCA for vessel enhancement in X-ray coronary angiograms”. Physics in Medicine and Biology 63 (July
2018). doi: 10.1088/1361-6560/aacddf.
[28] Binjie Qin et al. “Robust PCA Unrolling Network for Super-Resolution Vessel
Extraction in X-Ray Coronary Angiography”. IEEE Transactions on Medical
Imaging 41.11 (2022), pp. 3087–3098. doi: 10.1109/TMI.2022.3177626.
[29] Olaf Ronneberger, Philipp Fischer, and Thomas Brox. “U-Net: Convolutional
Networks for Biomedical Image Segmentation”. Medical Image Computing and
Computer-Assisted Intervention – MICCAI 2015. Ed. by Nassir Navab et al.
Cham: Springer International Publishing, 2015, pp. 234–241. isbn: 978-3-319-
24574-4.
[30] Ozan Oktay et al. “Attention U-Net: Learning Where to Look for the Pancreas”.
Medical Imaging with Deep Learning. 2018. url: https://openreview.
net/forum?id=Skft7cijM.
[31] National Taiwan University Medical Imaging. CAG biomedicalimages. 2021.
url: https://scidm.nchc.org.tw/dataset/cag.
[32] Fernando Cervantes-Sanchez et al. “Automatic Segmentation of Coronary Arteries in X-ray Angiograms using Multiscale Analysis and Artificial Neural
Networks”. Applied Sciences 9.24 (2019). issn: 2076-3417. doi: 10 . 3390 /
app9245507. url: https://www.mdpi.com/2076-3417/9/24/5507.
[33] Alexey Dosovitskiy et al. “An Image is Worth 16x16 Words: Transformers for
Image Recognition at Scale”. International Conference on Learning Representations. 2021. url: https://openreview.net/forum?id=YicbFdNTTy.
[34] Kaiming He et al. “Deep Residual Learning for Image Recognition”. 2016 IEEE
Conference on Computer Vision and Pattern Recognition (CVPR). 2016, pp. 770–
778. doi: 10.1109/CVPR.2016.90.
[35] Jia Deng et al. “ImageNet: A large-scale hierarchical image database”. 2009
IEEE Conference on Computer Vision and Pattern Recognition. 2009, pp. 248–
255. doi: 10.1109/CVPR.2009.5206848.
[36] Zongwei Zhou et al. “UNet++: A Nested U-Net Architecture for Medical Image Segmentation”. Deep Learning in Medical Image Analysis and Multimodal
Learning for Clinical Decision Support. Ed. by Danail Stoyanov et al. Cham:
Springer International Publishing, 2018, pp. 3–11. isbn: 978-3-030-00889-5.
[37] Liang-Chieh Chen et al. “Rethinking Atrous Convolution for Semantic Image
Segmentation”. ArXiv abs/1706.05587 (2017). url: https://api.semanticscholar.
org/CorpusID:22655199.
[38] Wentao Liu et al. “Multiscale Attention Aggregation Network for 2D Vessel Segmentation”. ICASSP 2022 - 2022 IEEE International Conference on Acoustics,
Speech and Signal Processing (ICASSP). 2022, pp. 1436–1440. doi: 10.1109/
ICASSP43922.2022.9747207.