Deep Feature Fusion with a Hybrid CNN–Transformer Framework for High-Fidelity Coronary Angiography Enhancement and Fine Vascular Structure Preservation
DOI:
https://doi.org/10.33411/IJIST/1998Keywords:
Convolutional neural networks, Deep learning, Coronary angiography, Image enhancement, Hybrid CNN–Transformer, Structural preservation, Feature fusion, Medical image restorationAbstract
Coronary Artery Disease (CAD) is a major cause of mortality worldwide, and X-ray coronary angiography remains an important imaging modality for evaluating coronary artery morphology and stenosis. However, low contrast, noise, motion blur, and compression artifacts can degrade angiographic images, obscure fine vascular structures, and reduce the reliability of subsequent image interpretation and automated analysis. To address these challenges, we propose a hybrid deep learning framework that integrates Convolutional Neural Networks (CNNs) with Transformer-based self-attention for high-fidelity coronary angiography enhancement. The CNN component extracts local features for noise suppression, edge refinement, and recovery of fine vessel details, while the Transformer component models long-range spatial dependencies to improve global anatomical continuity and structural coherence. The framework was trained in a supervised manner using high-quality images from the ARCADE dataset and their synthetically degraded counterparts. Quantitative evaluation demonstrated that the proposed Hybrid CNN–Transformer model achieved a PSNR of 36.33 dB and an SSIM of 0.9733, outperforming the Transformer-only model (34.43 dB, 0.9674 SSIM) by 1.90 dB in PSNR and 0.0059 in SSIM. The ablation analysis further showed that the proposed optimized hybrid configuration improved PSNR by 2.42 dB and SSIM by 0.0077 compared with the CNN-only configuration (33.91 dB, 0.9656 SSIM). The proposed model also achieved an MSE of 0.00023281 and an RMSE of 0.015258, representing the lowest reconstruction error among the evaluated methods. Qualitative assessment indicated improved vessel visibility, contrast, edge definition, thin-vessel representation, and preservation of vascular continuity. The results demonstrate that combining local convolutional feature extraction with global Transformer-based contextual modeling provides an effective approach for enhancing coronary angiographic images while preserving fine vascular structures, offering a promising preprocessing framework for subsequent computer-aided cardiovascular analysis.
References
S. Edpuganti, A. Shamim, V. H. Gangolli, R. A. D. K. N. W. Weerasekara, and A. Yellamilli, “Artificial Intelligence in Cardiovascular Imaging: Current Landscape, Clinical Impact, and Future Directions,” Discov. (Craiova, Rom., vol. 13, no. 1, p. e211, Jun. 2025, doi: 10.15190/D.2025.10.
S. Khan et al., “Sequence based model using deep neural network and hybrid features for identification of 5-hydroxymethylcytosine modification,” Sci. Reports 2024 141, vol. 14, no. 1, pp. 9116-, Apr. 2024, doi: 10.1038/s41598-024-59777-y.
S. Khan, I. Uddin, F. K. Alarfaj, and N. Almusallam, “A novel computational framework for tumor-specific T cell antigen identification using a deep neural network,” J. Comput. Mol. Des. 2026 401, vol. 40, no. 1, pp. 150-, Jun. 2026, doi: 10.1007/S10822-026-00861-Y.
I. Uddin et al., “Deep-m6Am: a deep learning model for identifying N6, 2′-O-Dimethyladenosine (m6Am) sites using hybrid features,” AIMS Bioeng. 2025 1145, vol. 12, no. 1, pp. 145–161, 2025, doi: 10.3934/BIOENG.2025006.
M. A. Gadiri, L. Pancaldi, and M. S. Sakar, “Flow-assisted and flow-driven endovascular medical instruments,” Recent Prog. Med. Miniat. Robot. From Bench to Bedside, pp. 1–19, Jan. 2025, doi: 10.1016/B978-0-443-13385-5.00004-0.
M. Alzamel, I. Uddin, S. Khan, N. Dilshad, and N. Ahmad, “Deep learning-based identification of N6-methyladenine sites via hybrid feature fusion and SHAP-driven feature selection,” Sci. Reports 2026 161, vol. 16, no. 1, pp. 22385-, May 2026, doi: 10.1038/s41598-026-50667-z.
I. Uddin, S. Noor, Y. A. Ali, S. Khan, and M. Al-Razgan, “A hybrid deep learning framework for accurate N6,2′-O-Dimethyladenosine site prediction,” Biophys. Chem., vol. 335, p. 107639, Aug. 2026, doi: 10.1016/J.BPC.2026.107639.
S. Asif et al., “Advancements and Prospects of Machine Learning in Medical Diagnostics: Unveiling the Future of Diagnostic Precision,” Arch. Comput. Methods Eng., vol. 32, no. 2, pp. 853–883, Mar. 2025, doi: 10.1007/S11831-024-10148-W/METRICS.
Q. M. Liao et al., “Computer-Aided Application in Medicine and Biomedicine,” Int. J. Comput. Intell. Syst. 2025 181, vol. 18, no. 1, pp. 221-, Aug. 2025, doi: 10.1007/S44196-025-00936-Y.
N. Dehlinger, J. P. Dillenseger, R. Pop, A. Hasiu, L. Bonnard, and I. Leonard-Lorant, “Towards standardised contrast injection protocols in diagnostic cerebral angiography for unruptured aneurysm evaluation: Addressing the current variability and technical challenges,” Radiography, vol. 32, no. 4, p. 103254, May 2026, doi: 10.1016/J.RADI.2025.103254.
J. Hu and E. L. E. Law, “A Comprehensive Review of Compressed Image Restoration and Compression Artifact Reduction Techniques,” IEEE Access, vol. 13, pp. 213824–213840, 2025, doi: 10.1109/ACCESS.2025.3644933.
I. Uddin et al., “A hybrid residue based sequential encoding mechanism with XGBoost improved ensemble model for identifying 5-hydroxymethylcytosine modifications,” Sci. Reports 2024 141, vol. 14, no. 1, pp. 20819-, Sep. 2024, doi: 10.1038/s41598-024-71568-z.
R. R. Kumar and R. Priyadarshi, “Denoising and segmentation in medical image analysis: A comprehensive review on machine learning and deep learning approaches,” Multimed. Tools Appl. 2024 8412, vol. 84, no. 12, pp. 10817–10875, May 2024, doi: 10.1007/S11042-024-19313-6.
E. Hassan, N. El-Rashidy, S. Elbedwehy, T. Abd El-Hafeez, A. Saber, and M. Y. Shams, “Exploring the frontiers of image super-resolution: a review of modern techniques and emerging applications,” Neural Comput. Appl. 2025 3722, vol. 37, no. 22, pp. 17913–17961, May 2025, doi: 10.1007/S00521-025-11331-1.
M. Aamir, Z. Rahman, N. Choudhry, J. Ahmed Bhutto, W. Ahmed Abro, and Z. Zhu, “From CNNs to Transformers: A Review of Evolving Deep Learning Architectures for Brain Tumor Classification,” IEEE Access, vol. 13, pp. 184918–184936, 2025, doi: 10.1109/ACCESS.2025.3625607.
C. Zhang, J. Wang, Y. Shi, B. Yin, and N. Ling, “A CNN-transformer hybrid network with selective fusion and dual attention for image super-resolution,” Multimed. Syst. 2025 312, vol. 31, no. 2, pp. 126-, Feb. 2025, doi: 10.1007/S00530-025-01711-X.
Steve Okyere-Gyamfi, Michael Asante, Kwame Ofosuhene Peasah, Yaw Marfo Missah, Vivian Akoto-Adjepong, “Contrast limited adaptive histogram equalization (CLAHE) and colour difference histogram (CDH) feature merging capsule network (CCFMCapsNet) for complex image recognition,” Plosone, vol. 20, no. 10, p. e0335393, 2025, [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/41171751/
B. Zhao, J. Peng, C. Chen, Y. Fan, K. Zhang, and Y. Zhang, “Diagnosis of Coronary Heart Disease Through Deep Learning-Based Segmentation and Localization in Computed Tomography Angiography,” IEEE Access, vol. 13, pp. 10177–10193, 2025, doi: 10.1109/ACCESS.2025.3528638.
C. Li, “A Review of Hybrid Models Combining Convolutional Neural Networks and Vision Transformers in Medical Image Processing,” Appl. Comput. Eng., vol. 222, no. 1, pp. 1–8, Jan. 2026, doi: 10.54254/2755-2721/2026.TJ31146.
Masroor Shah, Fazal Malik, Muhammad Suliman, Noor Rahman, Irfan Ullah, Sana Ullah, Romaan Khan, & Salman Alam, “Dark Data in Accident Prediction: Using AdaBoost and Random Forest for Improved Accuracy,” J. Comput. Biomed. Informatics, vol. 7, no. 2, 2024, [Online]. Available: https://jcbi.org/index.php/Main/article/view/531
“A Hierarchical Ensemble Deep Learning Framework for Angiographic Blood Vessel Segmentation and Coronary Artery Disease Classification | International Journal of Innovations in Science & Technology.” Accessed: Aug. 21, 2026. [Online]. Available: https://journal.50sea.com/index.php/IJIST/article/view/1983
I. Majid Mohammed and N. Ashidi Mat Isa, “Contrast Limited Adaptive Local Histogram Equalization Method for Poor Contrast Image Enhancement,” IEEE Access, vol. 13, pp. 62600–62632, 2025, doi: 10.1109/ACCESS.2025.3558506.
A. Paulauskaite-Taraseviciene, J. Siaulys, A. Jankauskas, and G. Jakuskaite, “A Robust Blood Vessel Segmentation Technique for Angiographic Images Employing Multi-Scale Filtering Approach,” J. Clin. Med. 2025, Vol. 14, Page 354, vol. 14, no. 2, p. 354, Jan. 2025, doi: 10.3390/JCM14020354.
M. Alkhodari, E. Alefisha, H. F. Jelinek, A. Kaabneh, and P. Liatsis, “Enhancing CCTA image quality: a review of deep learning approaches for advanced artifact correction and denoising,” Artif. Intell. Rev. 2025 5810, vol. 58, no. 10, pp. 331-, Aug. 2025, doi: 10.1007/S10462-025-11311-W.
C. Dong, C. C. Loy, K. He, and X. Tang, “Image Super-Resolution Using Deep Convolutional Networks,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 38, no. 2, pp. 295–307, Feb. 2016, doi: 10.1109/TPAMI.2015.2439281.
L. Gondara, “Medical Image Denoising Using Convolutional Denoising Autoencoders,” IEEE Int. Conf. Data Min. Work. ICDMW, vol. 0, pp. 241–246, Jul. 2016, doi: 10.1109/ICDMW.2016.0041.
K. Zhang, W. Zuo, Y. Chen, D. Meng, and L. Zhang, “Beyond a Gaussian denoiser: Residual learning of deep CNN for image denoising,” IEEE Trans. Image Process., vol. 26, no. 7, pp. 3142–3155, Jul. 2017, doi: 10.1109/TIP.2017.2662206.
F. Tatsugami et al., “Deep learning-based image restoration algorithm for coronary CT angiography,” Eur. Radiol., vol. 29, no. 10, pp. 5322–5329, Oct. 2019, doi: 10.1007/S00330-019-06183-Y.
L. Xiao, J. Song, X. Xie, and C. Fan, “Enhanced medical image segmentation using U-Net with residual connections and dual attention mechanism,” Eng. Appl. Artif. Intell., vol. 153, p. 110794, Aug. 2025, doi: 10.1016/J.ENGAPPAI.2025.110794.
X. Huang, Y. Ma, X. Mei, M. Sun, and S. Jia, “FastTrans-Net: Fusion network with shift window hierarchical attention in multi-level enhanced space for skin lesion segmentation,” Biomed. Signal Process. Control, vol. 112, p. 108675, Feb. 2026, doi: 10.1016/J.BSPC.2025.108675.
Y. Wen, A. Lai, B. Qian, W. Shi, and W. Cao, “Multi-weather Image Restoration via Histogram-Based Transformer Feature Enhancement,” Vis. Comput. 2025 4113, vol. 41, no. 13, pp. 11053–11067, Aug. 2025, doi: 10.1007/S00371-025-04085-3.
Z. Yu, F. Lee, and Q. Chen, “HCT-net: hybrid CNN-transformer model based on a neural architecture search network for medical image segmentation,” Appl. Intell. 2023 5317, vol. 53, no. 17, pp. 19990–20006, Mar. 2023, doi: 10.1007/S10489-023-04570-Z.
A. Senthil Anandhi and M. Jaiganesh, “An enhanced image restoration using deep learning and transformer based contextual optimization algorithm,” Sci. Reports 2025 151, vol. 15, no. 1, pp. 10324-, Mar. 2025, doi: 10.1038/s41598-025-94449-5.
I. Uddin, S. Khan, Y. A. Ali, N. Dilshad, N. Ahmad, and M. Al-Razgan, “Hybrid Deep Learning Framework for Accurate RNA m6Am Site Prediction,” 2025, doi: 10.2139/SSRN.5704553.
“Explainable AI-Driven Clinical Decision Support System for Cardiovascular Disease Prediction Using Hybrid Models | International Journal of Innovations in Science & Technology.” Accessed: Aug. 07, 2026. [Online]. Available: https://journal.50sea.com/index.php/IJIST/article/view/1962
S. Khan, I. Uddin, S. Noor, S. A. AlQahtani, and N. Ahmad, “N6-methyladenine identification using deep learning and discriminative feature integration,” BMC Med. Genomics 2025 181, vol. 18, no. 1, pp. 58-, Mar. 2025, doi: 10.1186/S12920-025-02131-6.
A. Ur Rahman, N. Iqbal, I. Uddin, and S. A. C. Bukhari, “Multifaceted Fusion of Glove and Additive Attention Models for Scalable Fake News Detection,” Proc. - 2025 IEEE Int. Conf. Futur. Mach. Learn. Data Sci. FMLDS 2025, pp. 363–368, 2025, doi: 10.1109/FMLDS67896.2025.00069.
I. Syed Ahmad Chan Bukhar., Nadeem Iqbal., Uddin., “Discriminative Sequence Features for DNA-Binding Protein Detection Using ML and DL Techniques,” reseach gate., pp. 461–421, doi: 10.1109/FMLDS67896.2025.0007.
“(PDF) Hybrid CNN-Transformer Framework for Structure-Preserving Enhancement of Coronary Angiography Images.” Accessed: Aug. 21, 2026. [Online]. Available: https://www.researchgate.net/publication/412164397_Hybrid_CNN-Transformer_Framework_for_Structure-Preserving_Enhancement_of_Coronary_Angiography_Images
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 50sea

This work is licensed under a Creative Commons Attribution 4.0 International License.


















