Preprocessing-Induced Ranking Reversal is Architecture-Dependent in Cross-Domain Pneumonia Detection
DOI:
https://doi.org/10.63158/journalisi.v8i5.1944Keywords:
pneumonia detection, convolutional neural network, image preprocessing, cross-domain generalization, chest X-rayAbstract
This study investigates how preprocessing selection influences the cross-domain generalization of convolutional neural networks (CNNs) for pneumonia detection and whether it changes performance rankings among architectures. Five preprocessing conditions—Baseline, Standard Deviation Normalization, Contrast Limited Adaptive Histogram Equalization (CLAHE), Histogram Equalization, and StdDev+CLAHE—were combined with three architectures: a simple CNN, ResNet-50, and EfficientNet-B0. The resulting 15 models were trained on the public Kaggle dataset containing 5,863 chest X-ray images and externally evaluated on 1,800 clinical chest X-rays collected in Central Java, Indonesia. Results showed that preprocessing rankings were architecture-dependent. The simple CNN achieved its highest external Recall with Histogram Equalization (0.9856), although Baseline ranked first internally. ResNet-50 was most stable with Standard Deviation Normalization, which ranked first internally (F1 = 0.9821) and externally (Recall = 0.9944). EfficientNet-B0 achieved the highest external Recall with CLAHE (0.9994), but its lower internal precision (0.77–0.87) indicated bias toward the pneumonia class. These findings demonstrate that high benchmark performance may not transfer directly to regional clinical settings because preprocessing interacts with scanner characteristics, data distribution, and network architecture. Therefore, each preprocessing–architecture pair should be validated on locally acquired images before clinical deployment.
Downloads
References
[1] “Pneumonia in children,” World Health Organization. Accessed: Aug. 28, 2026. [Online]. Available: https://www.who.int/news-room/fact-sheets/detail/pneumonia
[2] “Hasil Utama Survei Kesehatan Indonesia (SKI) 2023,” Badan Kebijakan Pembangunan Kesehatan, Kementerian Kesehatan Republik Indonesia. Accessed: Sep. 18, 2026. [Online]. Available: https://www.badankebijakan.kemkes.go.id/daftar-frequently-asked-question-seputar-hasil-utama-ski-2023/hasil-utama-ski-2023/
[3] S. A. Srigati, A. Munawir, and P. Wulandari, “Potensi kecerdasan buatan dalam peningkatan akurasi pembacaan hasil mamografi: Tinjauan sistematis dan meta-analisis,” Ganesha Medicina Journal, vol. 3, no. 1, pp. 65–71, Mar. 2023, doi: 10.23887/gm.v3i1.55575.
[4] F.-Z. Hamlili, M. Beladgham, M. Khelifi, and A. Bouida, “Transfer learning with Resnet-50 for detecting COVID-19 in chest X-ray images,” Indones. J. Electr. Eng. Comput. Sci., vol. 25, no. 3, pp. 1458–1468, Mar. 2022, doi: 10.11591/ijeecs.v25.i3.pp1458-1468.
[5] N. Y. Hegazy and M. S. Sawah, “An optimized EfficientNetB0 framework with CLAHE-based preprocessing for accurate multi-class chest X-ray classification,” Sci. Rep., vol. 16, no. 1, p. 10811, Mar. 2026, doi: 10.1038/s41598-026-42492-1.
[6] B. Guo et al., “The impact of scanner domain shift on deep learning performance in medical imaging: an experimental study,” Int. J. Comput. Assist. Radiol. Surg., early access, Apr. 2026, doi: 10.1007/s11548-026-03655-7.
[7] A. Giełczyk, A. Marciniak, M. Tarczewska, and Z. Lutowski, “Pre-processing methods in chest X-ray image classification,” PLoS One, vol. 17, no. 4, p. e0265949, Apr. 2022, doi: 10.1371/journal.pone.0265949.
[8] V. M, S. Rajkumar, M. V. Kalyan Reddy, and V. Janesh, “Detection of post COVID-pneumonia using histogram equalization, CLAHE deep learning techniques,” Intel. Artif., vol. 26, no. 72, pp. 137–145, Sep. 2023, doi: 10.4114/intartif.vol26iss72pp137-145.
[9] A. Raji and T. Olwal, “Performance evaluation of a hybrid histogram equalization-contrast limited adaptive histogram equalization technique for medical images enhancement,” Turk. J. Eng., vol. 10, no. 3, pp. 875–884, Jul. 2026, doi: 10.31127/tuje.1842375.
[10] A. N. bin Azhar, N. S. Sani, and L. L. X. Wei, “Enhancing COVID-19 detection in X-ray images through deep learning models with different image preprocessing techniques,” Int. J. Adv. Comput. Sci. Appl., vol. 16, no. 1, p. 61, 2025, doi: 10.14569/IJACSA.2025.0160161.
[11] V. Arya and T. Kumar, “Enhancing image for CNN-based diagnostic of pediatric pneumonia through chest radiographs,” Int. J. Adv. Comput. Sci. Appl., vol. 14, no. 2, p. 45, 2023, doi: 10.14569/IJACSA.2023.0140245.
[12] G. Siracusano, A. La Corte, A. G. Nucera, M. Gaeta, M. Chiappini, and G. Finocchio, “Effective processing pipeline PACE 2.0 for enhancing chest X-ray contrast and diagnostic interpretability,” Sci. Rep., vol. 13, no. 1, p. 22471, Dec. 2023, doi: 10.1038/s41598-023-49534-y.
[13] O. Dahmane, M. Khelifi, M. Beladgham, and I. Kadri, “Pneumonia detection based on transfer learning and a combination of VGG19 and a CNN Built from scratch,” Indones. J. Electr. Eng. Comput. Sci., vol. 24, no. 3, pp. 1469–1480, Dec. 2021, doi: 10.11591/ijeecs.v24.i3.pp1469-1480.
[14] D. S. Kermany et al., “Identifying medical diagnoses and treatable diseases by image-based deep learning,” Cell, vol. 172, no. 5, pp. 1122–1131.e9, Feb. 2018, doi: 10.1016/j.cell.2018.02.010.
[15] A. Musa, R. Prasad, and M. Hernandez, “Addressing cross-population domain shift in chest X-ray classification through supervised adversarial domain adaptation,” Sci. Rep., vol. 15, no. 1, p. 11383, Apr. 2025, doi: 10.1038/s41598-025-95390-3.
[16] P. M. Fernández-Miranda et al., “A retrospective study of deep learning generalization across two centers and multiple models of X-ray devices using COVID-19 chest-X rays,” Sci. Rep., vol. 14, no. 1, p. 14657, Jun. 2024, doi: 10.1038/s41598-024-64941-5.
[17] T. Kwon et al., “Diagnostic performance of artificial intelligence model for pneumonia from chest radiography,” PLoS One, vol. 16, no. 4, p. e0249399, Apr. 2021, doi: 10.1371/journal.pone.0249399.
[18] N. T. Vinh, L. T. Hien, H. M. Toan, and D. N. Toan, “Diagnosing tuberculosis from X-ray imaging based on contrast limited adaptive histogram equalization,” IAES Int. J. Artif. Intell., vol. 15, no. 3, pp. 2567–2580, Jun. 2026, doi: 10.11591/ijai.v15.i3.pp2567-2580.
[19] A. Shavkatovich Buriboev, A. Abduvaitov, and H. S. Jeon, “Binary classification of pneumonia in chest X-ray images using modified contrast-limited adaptive histogram equalization algorithm,” Sensors, vol. 25, no. 13, p. 3976, Jun. 2025, doi: 10.3390/s25133976.
[20] S. Karagol, K. N. Akpinar, and S. Genc, “Comparison of deep learning approaches in the classification of the chest X-ray,” J. Eng. Res., vol. 11, no. 3, p. 17121, Sep. 2023, doi: 10.36909/jer.17121.
[21] X. Huang et al., “External validation based on transfer learning for diagnosing atelectasis using portable chest X-rays,” Front. Med., vol. 9, p. 920040, Jul. 2022, doi: 10.3389/fmed.2022.920040.
[22] M. Abad, J. Casas-Roma, and F. Prados, “Generalizable disease detection using model ensemble on chest X-ray images,” Sci. Rep., vol. 14, no. 1, p. 5890, Mar. 2024, doi: 10.1038/s41598-024-56171-6.
[23] S. Gautam, M. M.-C. Hohne, S. Hansen, R. Jenssen, and M. Kampffmeyer, “Demonstrating the risk of imbalanced datasets in chest X-ray image-based diagnostics by prototypical relevance propagation,” in 2022 IEEE 19th International Symposium on Biomedical Imaging (ISBI), IEEE, Mar. 2022, pp. 1–5, doi: 10.1109/ISBI52829.2022.9761651.
[24] Z. Zhang, X. Zhang, K. Ichiji, I. Bukovský, and N. Homma, “How intra-source imbalanced datasets impact the performance of deep learning for COVID-19 diagnosis using chest X-ray images,” Sci. Rep., vol. 13, no. 1, p. 19049, Nov. 2023, doi: 10.1038/s41598-023-45368-w.
[25] D. Arias-Garzón, R. Tabares-Soto, J. Bernal-Salcedo, and G. A. Ruz, “Biases associated with database structure for COVID-19 detection in X-ray images,” Sci. Rep., vol. 13, no. 1, p. 3477, Mar. 2023, doi: 10.1038/s41598-023-30174-1.
[26] C. Ong Ly et al., “Shortcut learning in medical AI hinders generalization: method for estimating AI model generalization without external data,” NPJ Digit. Med., vol. 7, no. 1, p. 124, May 2024, doi: 10.1038/s41746-024-01118-4.
[27] B. G. Hill, F. L. Koback, and P. L. Schilling, “The risk of shortcutting in deep learning algorithms for medical imaging research,” Sci. Rep., vol. 14, no. 1, p. 29224, Nov. 2024, doi: 10.1038/s41598-024-79838-6.
[28] A. Brown, N. Tomasev, J. Freyberg, Y. Liu, A. Karthikesalingam, and J. Schrouff, “Detecting shortcut learning for fair medical AI using shortcut testing,” Nat. Commun., vol. 14, no. 1, p. 4314, Jul. 2023, doi: 10.1038/s41467-023-39902-7.
[29] A. K. Pathak, M. Gupta, and G. Jain, “Unmasking the Clever Hans effect in AI models: shortcut learning, spurious correlations, and the path toward robust intelligence,” Front. Artif. Intell., vol. 8, p. 1692454, Jan. 2026, doi: 10.3389/frai.2025.1692454.
[30] M. Haupt, E. Jolkver, and A. Schwerk, “Explainable AI for chest radiographs: sex-stratified fairness auditing in CNN-based pneumonia detection,” Sci. Rep., vol. 16, no. 1, p. 22709, Jul. 2026, doi: 10.1038/s41598-026-62761-3.
[31] R. Zhang, D. Griner, J. W. Garrett, Z. Qi, and G.-H. Chen, “Training certified detectives to track down the intrinsic shortcuts in COVID-19 chest X-ray data sets,” Sci. Rep., vol. 13, no. 1, p. 12690, Aug. 2023, doi: 10.1038/s41598-023-39855-3.
[32] A. Trivedi et al., “Deep learning models for COVID-19 chest X-ray classification: preventing shortcut learning using feature disentanglement,” PLoS One, vol. 17, no. 10, p. e0274098, Oct. 2022, doi: 10.1371/journal.pone.0274098.
[33] S. Ghimire, S. Kashyap, J. T. Wu, A. Karargyris, and M. Moradi, “Learning invariant feature representation to improve generalization across chest X-ray datasets,” in Machine Learning in Medical Imaging (MLMI 2020), Lecture Notes in Computer Science, vol. 12436. Cham: Springer, 2020, pp. 644–653, doi: 10.1007/978-3-030-59861-7_65.
[34] M. Zunaed, M. A. Haque, and T. Hasan, “Learning to generalize towards unseen domains via a content-aware style invariant model for disease detection from chest X-rays,” IEEE J. Biomed. Health Inform., vol. 28, no. 6, pp. 3626–3636, Jun. 2024, doi: 10.1109/JBHI.2024.3372999.
[35] S. Tayebi Arasteh, C. Kuhl, M.-J. Saehn, P. Isfort, D. Truhn, and S. Nebelung, “Enhancing domain generalization in the AI-based analysis of chest radiographs with federated learning,” Sci. Rep., vol. 13, no. 1, p. 22576, Dec. 2023, doi: 10.1038/s41598-023-49956-8.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Journal of Information Systems and Informatics
This work is licensed under a Creative Commons Attribution 4.0 International License.
Author's Declaration
- The Authors certify that they have read, understood, and agreed to the Journal of Information Systems and Informatics (JournalISI) submission guidelines, policies, and submission declaration. The submission has been prepared using the provided template.
- The Authors certify that all authors have approved the publication of this manuscript and that there is no conflict of interest.
- The Authors confirm that the manuscript is their original work, has not received prior publication, is not under consideration for publication elsewhere, and has not been previously published.
- The Authors confirm that all authors listed on the title page have contributed significantly to the work, have read the manuscript, attest to the validity and legitimacy of the data and its interpretation, and agree to its submission.
- The Authors confirm that the manuscript is not copied from or plagiarized from any other published work.
- The Authors declare that the manuscript will not be submitted for publication in any other journal or magazine until a decision is made by the journal editors.
- If the manuscript is finally accepted for publication, the Authors confirm that they will either proceed with publication immediately or withdraw the manuscript in accordance with the journal’s withdrawal policies.
- The Authors agree that, upon publication of the manuscript in this journal, they transfer copyright or assign exclusive rights to the publisher, including commercial rights
