An Optimized Solution for Accurate Disease Recognition and Classification of Gastrointestinal Endoscopic Images Based on a CNN-Transformer Hybrid Model
Vol. 28 (2026): 2026 2nd International Conference on Agricultural Sciences, Economics, Biomedical and Environmental Sciences (SEMBE 2026)
Received: 2026-07-18
Accepted: 2026-07-18
Published: 2026-07-18
Abstract
Accurate recognition and classification of gastrointestinal (GI) diseases through endoscopic images play a vital role in assisting early diagnosis, formulating effective treatment plans, and improving overall clinical decision-making. However, traditional convolutional neural networks (CNNs) often struggle to capture global contextual information in complex and highly variable GI images, frequently leading to suboptimal classification performance and limited clinical utility. In this paper, we propose an optimized and highly robust solution based on a novel hybrid CNN-Transformer architecture to systematically address these inherent challenges. Within this framework, the CNN component is primarily responsible for extracting fine-grained local texture features and spatial hierarchies, while the integrated Transformer module significantly enhances global semantic perception and long-range dependencies through advanced self-attention mechanisms. Furthermore, we introduce a sophisticated multi-scale feature fusion strategy to effectively integrate the complementary representations derived from both modules, ensuring a comprehensive understanding of the endoscopic visual data. To proactively tackle the pervasive issue of class imbalance and substantially improve model generalization across diverse patient demographics, advanced optimization techniques such as focal loss, label smoothing, and extensive domain augmentation are seamlessly incorporated. Extensive experiments conducted on a rigorously curated GI endoscopy dataset demonstrate that the proposed hybrid model significantly outperforms conventional CNN-based approaches across multiple evaluation metrics, particularly in terms of accuracy, recall, and overall robustness. Ultimately, our comprehensive solution offers a highly promising direction for building interpretable, efficient, and reliable computer-aided diagnostic systems tailored for clinical GI disease screening.
Keywords
References
[1] B. U. Eker and F. Z. Solak, "Robust, explainable, and statistically validated gastrointestinal image analysis using modern deep learning architectures," Signal, Image and Video Processing, vol. 19, no. 14, p. 1265, 2025.
[2] S. Ali, M. Dmitrieva, N. Ghatwary, S. Bano, G. Polat, A. Temizel, et al., "Deep learning for detection and segmentation of artefact and disease instances in gastrointestinal endoscopy," Medical Image Analysis, vol. 70, p. 102002, 2021.
[3] C. H. Kusters, T. J. Jaspers, T. G. Boers, M. R. Jong, J. B. Jukema, K. N. Fockens, et al., "Will Transformers change gastrointestinal endoscopic image analysis? A comparative analysis between CNNs and Transformers, in terms of performance, robustness and generalization," Medical Image Analysis, vol. 99, p. 103348, 2025.
[4] Z. Xu, Y. Miao, G. Chen, S. Liu, and H. Chen, "GLGFormer: Global Local Guidance Network for Mucosal Lesion Segmentation in Gastrointestinal Endoscopy Images," Journal of Imaging Informatics in Medicine, vol. 37, no. 6, pp. 2983–2995, 2024.
[5] M. A. Sheakh, M. S. Tahosin, M. J. Alam, and M. Begum, "Optimizing medical image analysis through MViTX on multiple datasets with explainable AI," Neural Computing and Applications, vol. 37, no. 14, pp. 8479–8507, 2025.
[6] T. T. H. Nguyen, X. H. Manh, T. T. Nguyen, H. V. Dao, and H. Vu, "EAGLE-Net: A hierarchical neural network for detecting anatomical landmarks in upper gastrointestinal endoscopy for clinical diagnosis," Healthcare Analytics, p. 100420, 2025.
[7] M. A. Elmagzoub, S. Kaur, S. Gupta, A. Rajab, K. D. Rajab, M. S. Al Reshan, et al., "Improving endoscopic image analysis: Attention mechanism integration in grid search fine-tuned transfer learning model for multi-class gastrointestinal disease classification," IEEE Access, vol. 12, pp. 80345–80358, 2024.
[8] T. Hossain, F. J. M. Shamrat, X. Zhou, I. Mahmud, M. S. A. Mazumder, S. Sharmin, and R. Gururajan, "Development of a multi-fusion convolutional neural network (MF-CNN) for enhanced gastrointestinal disease diagnosis in endoscopy image analysis," PeerJ Computer Science, vol. 10, p. e1950, 2024.
[9] T. Itoh, H. Kawahira, H. Nakashima, and N. Yata, "Deep learning analyzes Helicobacter pylori infection by upper gastrointestinal endoscopy images," Endoscopy International Open, vol. 6, no. 2, pp. E139–E144, 2018.
[10] T. Yan, P. K. Wong, and Y. Y. Qin, "Deep learning for diagnosis of precancerous lesions in upper gastrointestinal endoscopy: A review," World Journal of Gastroenterology, vol. 27, no. 20, p. 2531, 2021.
[11] J. K. Min, M. S. Kwak, and J. M. Cha, "Overview of deep learning in gastrointestinal endoscopy," Gut and Liver, vol. 13, no. 4, pp. 388, 2019.
[12] Q. He, S. Bano, O. F. Ahmad, B. Yang, X. Chen, P. Valdastri, et al., "Deep learning-based anatomical site classification for upper gastrointestinal endoscopy," International Journal of Computer Assisted Radiology and Surgery, vol. 15, no. 7, pp. 1085–1094, 2020.
[13] W. Du, N. Rao, D. Liu, H. Jiang, C. Luo, Z. Li, et al., "Review on the applications of deep learning in the analysis of gastrointestinal endoscopy images," IEEE Access, vol. 7, pp. 142053–142069, 2019.
[14] L. Wu, R. Shang, P. Sharma, W. Zhou, J. Liu, L. Yao, et al., "Effect of a deep learning-based system on the miss rate of gastric neoplasms during upper gastrointestinal endoscopy: a single-centre, tandem, randomised controlled trial," The Lancet Gastroenterology & Hepatology, vol. 6, no. 9, pp. 700–708, 2021.
[15] R. Pannala, K. Krishnan, J. Melson, M. A. Parsi, A. R. Schulman, S. Sullivan, et al., "Artificial intelligence in gastrointestinal endoscopy," VideoGIE, vol. 5, no. 12, pp. 598–613, 2020.
Copyright and License
Published in2026-07-18 15:34:29
DOI https://doi.org/10.70088/a3jaf441
Creative Commons
Copyright: © 2026 by the authors.
Submitted for possible open access
publication under the terms and
conditions of the Creative Commons
Attribution (CC BY) license
(https://creativecommons.org/license
s/by/4.0/).
Copyright
Copyright © The Author(s), 2026. Published by SEMBE 2026
Journal Information
- Vol. 28 (2026): 2026 2nd International Conference on Agricultural Sciences, Economics, Biomedical and Environmental Sciences (SEMBE 2026)
- 2026-07-18
- ISSN: (Print) 3078-770X/ (Online) 3078-7718
- Journal Homepage