Transfer learning as a regularized optimization : A mathematical and experimental model for medical image classification
*Manasi GhuleCorresponding authorghule.seema@gmail.comDepartment of Computer ScienceSNDT Women’s UniversityJuhu Campus, Mumbai, Maharashtra, 400049, IndiaView full profile → , Anita Chawareanita.chaware@computersc.sndt.ac.inDepartment of Computer ScienceSNDT Women’s UniversityJuhu Campus, Mumbai, Maharashtra, 400049, IndiaView full profile →
* Corresponding author · click or hover a name for details
- Received:
- 01 Dec 2025
- Published Online:
- 30 Sep 2026
- Article type:
- Research Article
- Language:
- EN
- Article no.:
- JIOS-2355
- Pages:
- 3153–3166
Abstract
Transfer learning has emerged as a pivotal technique for biomedical image classification, where labeled data is limited and computation cost of model training is very high. By leveraging pre-trained neural networks, transfer learning accelerates convergence and enhances model precision on domain-specific tasks. The efficiency and performance of transfer learning is governed by various factors that need to be fine-tuned. Despite its wide success, the mathematical understanding of transfer learning behavior remains limited. In this paper, we systematically studied the mathematical aspect of transfer learning as a regularized optimization problem, where adaptation to a target domain is balanced against deviation from pre-trained parameters, which plays an important role in classification of datasets including small and imbalanced datasets invariably found in medical diagnosis. We validated this model using controlled regularization across a simulated neural network and three models of deep convolutional networks pre-trained on two different domains for classification of ultrasound scans used in diagnosis of polycystic ovary syndrome. Our experimental results suggest that regularization improves generalization and preserves inductive bias, due to a principled adaptation-retention tradeoff, thereby bridging mathematical modeling and practical implementation of transfer learning in domain-adaptive medical imaging applications.
Keywords
Subject Classifications
References
[1] A. Shah, W. Patel, and H. Koyuncu, “Empowering healthcare innovation: IoT-enabled smart systems and deep learning for enhanced diabetic retinopathy in the telehealth landscape,” J. Interdiscip. Math., vol. 27, no. 2, pp. 355–367 (2024), doi: 10.47974/JIM-1836.
[2] H. Zhang and Y. Qie, “Applying deep learning to medical imaging: A review,” Appl. Sci., vol. 13, no. 18, p. 10521 (2023), doi: 10.3390/app131810521.
[3] D. P. Sahoo, M. Rout, P. K. Mallick, and S. R. Samanta, “Comparative analysis of medical images using transfer learning based deep learning models,” 2022 International Conference on Advancements in Smart, Secure and Intelligent Computing (ASSIC), Bhubaneswar, India, pp. 1-8 (2022), doi: 10.1109/ASSIC55218.2022.10088373.
[4] F. Zhuang, Q. Qi, K. Duan, D. Xi, Y. Zhu, H. Zhu, H. Xiong, and Q. He, “A comprehensive survey on transfer learning,” arXiv:1911.02685 (2020), doi: 10.48550/arXiv.1911.02685.
[5] V. Gunturu, N. Maiti, B. Toure, P. Kunekar, S. B. Banu, and D. Sahaya Lenin, “Transfer learning in biomedical image classification,” in Proc. Int. Conf. Adv. Comput., Commun. Appl. Informatics (ACCAI), pp. 1–5 (2024), doi: 10.1109/ACCAI61061.2024.10601862.
[6] A. Seethalakshmy, T. Tamilvizhi, K. Naga Sowjanya, and B. Bala, “Deep learning based computed tomography image classification of COVID-19 patients,” J. Interdiscip. Math., vol. 26, no. 3, pp. 371–381 (2023), doi: 10.47974/JIM-1668.
[7] H. Cao, H. Gu, and X. Guo, “Feasibility of transfer learning: A mathematical framework,” arXiv:2305.12985 (2023), doi: 10.48550/arXiv.2305.12985.
[8] N. Mamatha and B. Shetty, “Machine learning mathematical optimization techniques for enhancing model efficiency and convergence,” World J. Adv. Res. Rev., vol. 10, no. 3, pp. 471–481 (2021), doi: 10.30574/wjarr.2021.10.3.0235.
[9] M. Long, J. Wang, G. Ding, S. J. Pan, and P. S. Yu, “Adaptation regularization: A general framework for transfer learning,” IEEE Trans. Knowl. Data Eng., vol. 26, no. 5, pp. 1076–1089 (2014), doi: 10.1109/TKDE.2013.111.
[10] X. Li, Y. Grandvalet, and F. Davoine, “A baseline regularization scheme for transfer learning with convolutional neural networks,” Pattern Recognit., vol. 98, p. 107049 (2020), doi: 10.1016/j.patcog.2019.107049.
[11] S. J. Pan and Q. Yang, “A survey on transfer learning,” IEEE Trans. Knowl. Data Eng., vol. 22, no. 10, pp. 1345–1359 (2010), doi: 10.1109/TKDE.2009.191.
[12] I. Goodfellow, Y. Bengio, and A. Courville, Deep Learning. Cambridge, MA, USA: MIT Press (2016). [Online]. Available: http://www.deeplearningbook.org
[13] S. J. Pan, X. Ni, J.-T. Sun, Q. Yang, and Z. Chen, “Cross-domain sentiment classification via spectral feature alignment,” in Proc. 19th Int. Conf. World Wide Web (WWW ’10), New York, USA: Association for Computing Machinery, pp. 751–760 (2010), doi: 10.1145/1772690.1772767.
[14] W. Zhang, P. Zhang, B. Zhang, X. Wang, and D. Wang, “A collaborative transfer learning framework for cross-domain recommendation,” arXiv:2306.16425 (2023), doi: 10.48550/arXiv.2306.16425.
[15] T. Cody and P. A. Beling, “A systems theory of transfer learning,” arXiv:2107.01196 (2021), doi: 10.48550/arXiv.2107.01196.
[16] P. Patil, S. Charkha, and S. V. Kakade, “Transfer learning optimization in medical imaging enhancing diagnostic accuracy with limited data sets,” J. Inf. Optim. Sci., vol. 46, no. 4-A, pp. 1105–1116 (2025), doi: 10.47974/JIOS-1895.
[17] F. Yu, X. Xiu, and Y. Li, “A survey on deep transfer learning and beyond,” Mathematics, vol. 10, no. 19, p. 3619 (2022), doi: 10.3390/math10193619.
[18] Y. Tan and X. Xiang, “Illustrating Ben-David et al.’s theory of learning from different domains,” (2021), doi: 10.13140/RG.2.2.26939.13607.
[19] K. P. Murphy, Machine Learning: A Probabilistic Perspective. Cambridge, MA, USA: MIT Press (2012).
[20] X. Li, “Regularization schemes for transfer learning with convolutional networks,” Ph.D. dissertation, Université de Technologie de Compiègne (2019). [Online]. Available: https://theses.hal.science/tel-02439193.
[21] J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei-Fei, “ImageNet: A large-scale hierarchical image database,” in 2009 IEEE Conference on Computer Vision and Pattern Recognition, pp. 248–255 (2009), doi: 10.1109/CVPR.2009.5206848.
[22] J. Yang, R. Shi, D. Wei, Z. Liu, L. Zhao, B. Ke, H. Pfister, B. Ni, and M. Xu, “MedMNIST v2: A large-scale lightweight benchmark for 2D and 3D biomedical image classification,” Sci. Data, vol. 10, no. 1, p. 41 (2023), doi: 10.1038/s41597-022-01721-8.
[23] G. Peyré and M. Cuturi, “Computational optimal transport,” arXiv:1803.00567 (2020), doi: 10.48550/arXiv.1803.00567.
[24] W. Li, J. Lu, and L. Wang, “Fisher information regularization schemes for Wasserstein gradient flows,” J. Comput. Phys., vol. 416, p. 109449 (2020), doi: 10.1016/j.jcp.2020.109449.
