Deep Transfer Learning with Customized Convolutional Neural Networks for Radiological Image Classification
DOI:
https://doi.org/10.29284/gtd7vf52Keywords:
Radiology, ResNet50, Gradient-weighted Class Activation Mapping, CNNAbstract
Radiological image interpretation is fundamental to modern clinical diagnosis, yet remains challenging due to high data dimensionality, complex anatomical variability, and the limited availability of labeled datasets. This study proposes a deep learning framework that integrates a customized convolutional neural network (CNN) with ResNet50- based transfer learning to improve diagnostic accuracy and robustness in radiology. The bespoke CNN is designed to capture hierarchical spatial representations from medical images, while ResNet50 leverages pretrained weights to accelerate convergence, preserve generic visual features, and enhance generalization to domain-specific radiological patterns. Extensive experiments conducted on multiple radiology datasets demonstrate that the proposed approach consistently outperforms CNNs trained from scratch, achieving accuracy exceeding 91% alongside notable improvements in Precision, Recall, F1-score, and Area Under the Curve (AUC). Gradient-weighted Class Activation Mapping (Grad-CAM) further reveals that the model focuses on clinically relevant regions such as lesions and abnormal tissue structures, thereby providing interpretable predictions and strengthening clinical trust. The results confirm that transfer learning significantly mitigates overfitting, stabilizes model training, and delivers reliable performance in limited-data scenarios commonly encountered in medical imaging. In addition, the proposed framework exhibits faster convergence and improved consistency across heterogeneous datasets, highlighting its robustness and adaptability. Overall, this work presents a scalable and explainable solution for automated radiological diagnosis, demonstrating how the integration of pretrained deep architectures with task-specific CNN design can advance AI-assisted clinical decision support and contribute toward more accurate, efficient, and reliable medical image analysis.
References
[1] Senthil Pandi, S., Kumaragurubaran, T., Jaeyalakshmi, M., & Sheriff, M. A. (2024). A Comparative Analysis on the
Prediction of Heart Failure using Machine Learning Algorithms. In Proceedings - 2024 5th International
Conference on Mobile Computing and Sustainable Informatics, ICMCSI 2024 (pp. 206–211). Institute of Electrical
and Electronics Engineers Inc. https://doi.org/10.1109/ICMCSI61536.2024.0003
[2] Saravana Gokul, G., Senthil Pandi, S., Deepak Kumar, K., Sheriff, M. A., & Krishnamurthy, R. (2024). Optimizing
Event Management Application through Data Analytics Driver and Dynamic Exploration. In 2nd IEEE
International Conference on Advances in Information Technology, ICAIT 2024 - Proceedings. Institute of Electrical
and Electronics Engineers Inc. https://doi.org/10.1109/ICAIT61638.2024.10690407
[3] Jesuraj, Y. A., & Hemalatha, K. (2020). A prototype model of unmanned automatic level crossing system using
piezoelectric sensor. Microprocessors and Microsystems, 79. https://doi.org/10.1016/j.micpro.2020.103265
[4] Mohamed Yacin Sikkandar, K. Hemalatha, M. Subashree, S. Srinivasan, Seifedine Kadry, Jungeun Kim, Keejun
Han, Leveraging Multimodal Ensemble Fusion-Based Deep Learning for COVID-19 on Chest Radiographs,
Computer Systems Science and Engineering, 47(1),2023, 873-889, https://do i.org/10 .32604/c
sse.2023.035730.
[5] Afrah Fathima, S., Sheriff, M. A., Vyshnavi, A. R., & Jananee, V. (2023). INTEGRATING FUZZY LOGIC WITH SUPPLY
CHAIN MANAGEMENT. In 2023 IEEE International Conference on Research Methodologies in Knowledge
Management, Artificial Intelligence and Telecommunication Engineering, RMKMATE 2023. Institute of Electrical
and Electronics Engineers Inc. https://doi.org/10.1109/RMKMATE59243.2023.10368961
[6] Senthil Pandi, S., Kumar, P., Nandha Kumar, P., & Karthick, R. (2025). Machine Unlearning for Grid SearchCV.
In 2025 International Conference on Computing Technologies and Data Communication, ICCTDC 2025. Institute
of Electrical and Electronics Engineers Inc. https://doi.org/10.1109/ICCTDC64446.2025.11158784
[7] Salve, S. S., & Narote, S. P. (2024). Performance evaluation of efficient segmentation and classification based iris
recognition using sheaf attention network. Journal of Visual Communication and Image Representation, 103.
https://doi.org/10.1016/j.jvcir.2024.104262
[8] Senthil Pandi, S., Bala Subramanian, C., Praveen, S. D., & Ezhil Dyana, V. M. (2024). Improving Skin Lesion
Diagnosis Employing Attention-Based LSTM Networks. In the 2024 International Conference on Integration of
Emerging Technologies for the Digital World, ICIETDW 2024. Institute of Electrical and Electronics Engineers
Inc. https://doi.org/10.1109/ICIETDW61607.2024.10939196
[9] M. Thilagaraj, N. Arunkumar, S. Ramkumar, S. Hariharasitaraman, Electrooculogram signal identification for
elderly disabled using Elman network, Microprocessors and Microsystems, Volume 82, 2021, 103811, ISSN
0141-9331, https://doi.org/10.1016/j.micpro.2020.103811.
[10] Arockia Jesuraj Yagappan, Hemalatha Karuppiah, Maragatharajan Muthusamy, Sathish Kumar Kannaiah,
Optimizing lung cancer classification using transformer and Gooseneck Barnacle Optimization, Expert Systems
with Applications, Volume 282, 2025, 127413, ISSN 0957-4174, https://doi.org/10.1016/j.eswa.2025.127413.
[11] Jeba Prasanna Idas, S., Hemalatha, K., Naveenkumar, J. et al. Recent trends on mammogram breast density
analysis using deep learning models: neoteric review. Artif Intell Rev 58, 240 (2025).
https://doi.org/10.1007/s10462-025-11232-8
[12] Naga Pavan, M. E., Reddy, U. B. K., Ghantasala, C. M., Devi, M. M. Y., & Nesarani, A. (2023). Enhanced Skin Cancer
Classification with AlexNet and Transfer Learning. In International Conference on Sustainable Communication
Networks and Application, ICSCNA 2023 - Proceedings (pp. 1020–1024). Institute of Electrical and Electronics
Engineers Inc. https://doi.org/10.1109/ICSCNA58489.2023.10370110
[13] S. S. Salve. (2024). Efficient Iris Recognition Using Hybrid Preprocessing and Sheaf Attention Network-Based
Segmentation and Classification. Educational Administration: Theory and Practice, 30(1), 8319–8329.
https://doi.org/10.53555/kuey.v30i1.11437
[14] Singh, A. R., Athisayamani, S., Hariharasitaraman, H. S., Karim, F. K., Varela-Aldas, J., & Mostafa, S. M. (2025).
Depth-Enhanced Tumor Detection Framework for Breast Histopathology Images by Integrating Adaptive Multi-
Scale Fusion, Semantic Depth Calibration, and Boundary-Guided Detection. IEEE Access, 13, 58735–58745.
https://doi.org/10.1109/ACCESS.2025.3554342
[15] Sushilkumar S. (2023). Hybrid Image Enhancement and Dense Segmentation for Robust Iris Recognition
Using Attention Mechanism. Journal for ReAttach Therapy and Developmental Diversities, 6(10s(2), 2398–2408.
https://doi.org/10.53555/jrtdd.v6i10s(2).3846
[16] Alex David, S., Begum, A., Kalilulah, S. I., Ruth Naveena, N., Rajathi, K., & Hemalatha, D. (2024). Deep Learning
Advances in Brain Tumor Classification: Leveraging VGG16 and MobileNetV2 for Accurate MRI Diagnostics.
In 4th International Conference on Power, Energy, Control and Transmission Systems: Harnessing Power and
Energy for an Affordable Electrification of India, ICPECTS 2024. Institute of Electrical and Electronics Engineers
Downloads
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution 4.0 International License.
This work is licensed under a Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
