The Clinical Utility of Compressed ResNetGhost vs. ResNet-50: A Comparative Study for COVID-19 CT Diagnosis
Corresponding author: Kadhim Aseel Nadhum
Abstract
ResNetGhost constitutes a stage-selective optimization of ResNet-50, strategically integrating Ghost modules into Stages 3–4 to minimize computational redundancy while maintaining diagnostic fidelity in classifying COVID-19 CT severity. This architecture was validated on a multi-center cohort of 577 patients to address potential domain shifts through a rigorous patient-level split. In an independent test set (n = 164), ResNetGhost demonstrated superior discriminative power with an ROC-AUC of 0.991 and a verified accuracy of 97.56% (160/164 cases). Furthermore, this method effectively resolved systematic over-confidence, achieving a high-fidelity ECE of 0.0175. McNemar's test (p = 0.1573) confirmed that the 62.29% parameter reduction and 30% FLOPs decrease did not induce statistically significant degradation in diagnostic integrity compared to the baseline. Consequently, ResNetGhost offers a robust, well-calibrated solution optimized for resource-constrained clinical environments.
Keywords:
COVID-19, chest CT, severity classification, ResNet-50, Ghost modules, model compressionReferences
G. Rasul et al., "Socio-Economic Implications of COVID-19 Pandemic in South Asia: Emerging Risks and Growing Challenges," Frontiers in Sociology, vol. 6, Feb. 2021.
K. A. Nadhum, S. M. Sam, and S. Usman, "Prediction Model Using Deep Learning for Lung Illness Severity Among Covid-19 Patients in Iraq," in 2024 5th International Conference on Smart Sensors and Application (ICSSA), Sept. 2024, pp. 1–6.
M. Chhabra and R. Kumar, "An Efficient ResNet-50 based Intelligent Deep Learning Model to Predict Pneumonia from Medical Images," in 2022 International Conference on Sustainable Computing and Data Communication Systems (ICSCDS), Apr. 2022, pp. 1714–1721.
D. Suganya. and R. Kalpana., "Prognosticating various acute covid lung disorders from COVID-19 patient using chest CT Images," Engineering Applications of Artificial Intelligence, vol. 119, Mar. 2023, Art. no. 105820.
M. A. Mezher, S. B. Alrifai, and W. M. Raoof, "Analysis of Proinflammatory Cytokines in COVID-19 Patients in Baghdad, Iraq," Archives of Razi Institute, vol. 78, no. 1, pp. 305–313, Feb. 2023.
J. F. Abdulkareem and H. K. Aljobouri, "Chest CT images analysis with deep learning algorithms for COVID-19 diagnostic for Iraqi center," AIP Conference Proceedings, vol. 2414, no. 1, Feb. 2023, Art. no. 060004.
T. Liang, J. Glossner, L. Wang, S. Shi, and X. Zhang, "Pruning and quantization for deep neural network acceleration: A survey," Neurocomputing, vol. 461, pp. 370–403, Oct. 2021.
Y. He and L. Xiao, "Structured Pruning for Deep Convolutional Neural Networks: A Survey," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 46, no. 5, pp. 2900–2919, Feb. 2024.
X. Jiang, X. Bian, and C. Guo, "Ghost-Stereo: GhostNet-based Cost Volume Enhancement and Aggregation for Stereo Matching Networks." arXiv, May 23, 2024.
H. Hussain, P. S. Tamizharasan, and P. K. Yadav, "LCRM: Layer-Wise Complexity Reduction Method for CNN Model Optimization on End Devices," IEEE Access, vol. 11, pp. 66838–66857, 2023.
Z. Wang and T. Li, "A Lightweight CNN Model Based on GhostNet," Computational Intelligence and Neuroscience, vol. 2022, pp. 1–12, July 2022.
S. H. Khan, "COVID-19 Detection and Analysis From Lung CT Images using Novel Channel Boosted CNNs." arXiv, Sept. 01, 2022.
K. A. Nadhum, S. B. M. Sam, and S. B. Usman, "Optimizing ResNet50 for Medical Image Classification: A Comparative Study of Ghost Modules, Pruning, and Knowledge Distillation," Engineering, Technology & Applied Science Research, vol. 15, no. 6, pp. 28544–28549, Dec. 2025.
Y. Fu, Y. Lei, T. Wang, W. J. Curran, T. Liu, and X. Yang, "Deep learning in medical image registration: a review," Physics in Medicine & Biology, vol. 65, no. 20, Oct. 2020, Art. no. 20TR01.
Y. Tang, K. Han, J. Guo, C. Xu, C. Xu, and Y. Wang, "GhostNetV2: Enhance Cheap Operation with Long-Range Attention," Advances in Neural Information Processing Systems, vol. 35, pp. 9969–9982, Dec. 2022.
C. Shorten, T. M. Khoshgoftaar, and B. Furht, "Text Data Augmentation for Deep Learning," Journal of Big Data, vol. 8, no. 1, July 2021, Art. no. 101.
O. A. Adedokun and W. D. Burgess, "Analysis of Paired Dichotomous Data: A Gentle Introduction to the McNemar Test in SPSS," Journal of MultiDisciplinary Evaluation, vol. 8, no. 17, pp. 125–131, Jan. 2012.
H. K. Hamarashid, "Utilizing Statistical Tests for Comparing Machine Learning Algorithms," Kurdistan Journal of Applied Research, vol. 6, no. 1, pp. 69–74, July 2021.
Z. Zhang et al., "Decision curve analysis: a technical note," Annals of Translational Medicine, vol. 6, no. 15, Aug. 2018, Art. no. 308.
R. Vasilev and A. D'yakonov, "Calibration of Neural Networks." arXiv, 2023.
Downloads
How to Cite
License
Copyright (c) 2026 Kadhim Aseel Nadhum, Suriani Binti Mohd Sam, Sahnius Bt. Usman
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain the copyright and grant the journal the right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) after its publication in ETASR with an acknowledgement of its initial publication in this journal.
