Dynamic Uncertainty-Aware Adaptive Subspace Fusion Network for Robust Multimodal Medical Image Classification
Abstract
Multimodal medical image classification leverages complementary information from multiple imaging modalities to improve diagnostic accuracy and clinical decision-making. However, most existing multimodal fusion approaches rely on deterministic low-rank constraints and assume equal importance across all modalities. Such assumptions significantly limit flexibility, robustness, and interpretability, particularly in real-world clinical scenarios where modality data may be noisy, incomplete, or partially missing. To address these challenges, this work proposes a Dynamic Uncertainty-Aware Adaptive Subspace Fusion Network (DUA-SFNet) for robust multimodal medical image classification. The core of the proposed framework is a rank-learning adaptive-rank tensor decomposition module that dynamically adjusts subspace dimensionality according to the intrinsic complexity of the input data. This adaptive mechanism effectively reduces feature redundancy while preserving the highly discriminative information essential for accurate classification. In addition, DUA-SFNet incorporates a modality uncertainty estimation scheme to explicitly quantify the reliability and trustworthiness of each modality. By assigning uncertainty-aware weights during the fusion process, the framework can suppress unreliable or noisy modalities while emphasizing more informative ones, thereby improving resilience under adverse data conditions. Furthermore, a hierarchical adaptive attention strategy is employed to jointly model intra-subspace feature interactions and inter-modality dependencies. This design enhances feature representation capability while offering improved clinical interpretability by revealing how different modalities and subspaces contribute to the final decision. Extensive experiments conducted on multiple public and self-organized multimodal medical image datasets demonstrate that DUA-SFNet consistently outperforms state-of-the-art methods, achieving classification accuracy improvements of 3.8–6.2% and F1-score gains of 4.1–7.5%. Overall, DUA-SFNet provides an interpretable, uncertainty-aware, and adaptive solution for next-generation multimodal medical image analysis.
Downloads
References
Zhang, X., Xie, Z., Yu, H., Wang, Q., Wang, P., & Wang, W. (2024, October). Enhancing Adaptive Deep Networks for Image Classification via Uncertainty-aware Decision Fusion. In Proceedings of the 32nd ACM International Conference on Multimedia (pp. 8595-8603). https://doi.org/10.1145/3664647.3681368
Zhu, Q., Zheng, C., Zhang, Z., Shao, W., & Zhang, D. (2023). Dynamic confidence-aware multi-modal emotion recognition. IEEE Transactions on Affective Computing, 15(3), 1358-1370. https://doi.org/10.1109/TAFFC.2023.3340924
Guo, J., Cheng, Y., He, W., Zhang, Y., Feng, R., & Zhang, X. (2025, April). Uncertainty-Aware Dynamic Fusion for Multimodal Clinical Prediction Tasks. In ICASSP 2025-2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (pp. 1-5). IEEE. https://doi.org/10.1109/ICASSP49660.2025.10889595
Du, L., Liu, C., Wei, R., & Chen, J. (2023). Uncertainty-aware dynamic integration for multi-omics classification of tumors. Journal of Cancer Research and Clinical Oncology, 149(7), 3301-3312. https://doi.org/10.1007/s00432-022-04219-3
Shao, Z., Wang, H., Cai, Y., Chen, L., & Li, Y. (2025). UA-Fusion: Uncertainty-Aware Multimodal Data Fusion Framework for 3D Object Detection of Autonomous Vehicles. IEEE Transactions on Instrumentation and Measurement. https://doi.org/10.1109/TIM.2025.3548184
E. Munari, A. Scarpa, L. Cima, M. Pozzi, F. Pagni, F. Vasuri, S. Marletta, A.P. Dei Tos, A. Eccher, Cutting-edge technology and automation in the pathology laboratory, Virchows Arch. 484 (4) (2024) 555–566. https://doi.org/10.1007/s00428-023-03637-z
Hao, D., Meng, M., Gao, Y., Lou, X., & Kong, W. (2025). Step-wise Prompting Meets Uncertainty-Aware Dynamic Fusion for Robust EEG-Visual Emotion Recognition. IEEE Transactions on Affective Computing. https://doi.org/10.1109/TAFFC.2025.3632304
Wen, J., Long, J., Lu, X., Liu, C., Fang, X., & Xu, Y. (2025). Partial multiview incomplete multilabel learning via uncertainty-driven reliable dynamic fusion. IEEE Transactions on Pattern Analysis and Machine Intelligence. https://doi.org/10.1109/TPAMI.2025.3603677
Xu, W., Liu, X., Wang, J., Zhang, F., Hu, D., & Zong, L. (2025). UAMRL: multi-granularity uncertainty-aware multimodal representation learning for drug-target affinity prediction. Bioinformatics, 41(10), btaf512. https://doi.org/10.1093/bioinformatics/btaf512
Li, F. J., Zhang, C. Y., & Chen, C. P. (2025). Robust Decision-Making Method for Autonomous Driving Based on Multi-Modal Fusion and Uncertainty Modeling. IEEE Transactions on Vehicular Technology. https://doi.org/10.1109/TVT.2025.3613839.
Abdusalomov, A., Umirzakova, S., Boymatov, E., Zaripova, D., Kamalov, S., Temirov, Z., ... & Whangbo, T. K. (2025). A Human-Centric, Uncertainty-Aware Event-Fused AI Network for Robust Face Recognition in Adverse Conditions. Applied Sciences, 15(13), 7381. https://doi.org/10.3390/app15137381
Xie, W., Lu, X., Liu, Y., Long, J., Zhang, B., Zhao, S., & Wen, J. (2024, October). Uncertainty-aware pseudo-labeling and dual graph driven network for incomplete multi-view multi-label classification. In Proceedings of the 32nd ACM international conference on multimedia (pp. 6656-6665). https://doi.org/10.1145/3664647.3680932
Koishiyeva, D., Mukhammejanova, D., Kang, J. W., & Mukasheva, A. (2025). A Review of Deep Learning Approaches Based on Segment Anything Model for Medical Image Segmentation. Bioengineering, 12(12), 1312. https://doi.org/10.3390/bioengineering12121312
Wang, Y., Song, D., Wang, W., Rao, S., Wang, X., & Wang, M. (2022). Self-supervised learning and semi-supervised learning for multi-sequence medical image classification. Neurocomputing, 513, 383-394. https://doi.org/10.1016/j.neucom.2022.09.097
Jiang, J., Zhou, Q., Chen, N., He, H., Zhang, J., & He, C. Davlmf-Seg: Vision-Language Model Guided Latent Frequency-Aware Diffusion for Semi-Supervised Medical Image Segmentation. Available at SSRN 5387264. http://dx.doi.org/10.2139/ssrn.5387264
S. Iqbal, A.N. Qureshi, M. Alhussein, K. Aurangzeb, M. Zubair, A. Hussain, A novel reciprocal domain adaptation neural network for enhanced diagnosis of chronic kidney disease, Expert Syst. 42 (2) (2025) e13825. https://doi.org/10.1111/exsy.13825.
A. Shabbir, M. Zubair, Interpretable deep learning classifier using explainable AI for non-small cell lung cancer, in: 2024 Horizons of Information Technology and Engineering, HITE, IEEE, 2024, pp. 1–6. https://doi.org/10.1109/HITE63532.2024.10777248
E. Bercovich, M.C. Javitt, Medical imaging: from roentgen to the digital revolution, and beyond, Rambam Maimonides Med. J. 9 (4) (2018). https://doi.org/10.5041/RMMJ.10355
Lu, Y., Zhao, Y., Chen, X., & Guo, X. (2022). A Novel U‐Net Based Deep Learning Method for 3D Cardiovascular MRI Segmentation. Computational Intelligence and Neuroscience, 2022(1), 4103524. https://doi.org/10.1155/2022/4103524
Suganyadevi, S., Pershiya, A. S., Balasamy, K., et al. “Deep learning based alzheimer disease diagnosis: A comprehensive review”. SN Computer Science, Vol.5 no.4, pp.391, 2024, https://doi.org/10.1007/s42979-024-02743-2.
Balasamy, K., Krishnaraj, N., & Vijayalakshmi, K. “An adaptive neuro-fuzzy based region selection and authenticating medical image through watermarking for secure communication”, Wireless Personal Communications, Vol.122, no.3, pp. 2817–2837, 2021, https://doi.org/10.1007/s11277-021-09031-9.
Suganyadevi, S., & Seethalakshmi, V. “CVD-HNet: Classifying Pneumonia and COVID-19 in Chest X-ray Images Using Deep Network”. Wireless Personal Communications, Vol.126, no. 4, pp.3279–3303, 2022, https://doi.org/10.1007/s11277-022-09864-y
Balasamy, K., & Suganyadevi, S. “Multi-dimensional fuzzy based diabetic retinopathy detection in retinal images through deep CNN method”. Multimedia Tools and Applications, Vol 83, no. 5, pp.1–23. 2024, https://doi.org/10.1007/s11042-024-19798-1
Balasamy, K., Seethalakshmi, V. & Suganyadevi, S. Medical Image Analysis Through Deep Learning Techniques: A Comprehensive Survey. Wireless Pers Commun, 137, 1685–1714 (2024). https://doi.org/10.1007/s11277-024-11428-1
Suganyadevi, S., Seethalakshmi, V. Deep recurrent learning based qualified sequence segment analytical model (QS2AM) for infectious disease detection using CT images. Evolving Systems, 15, 505–521 (2024). https://doi.org/10.1007/s12530-023-09554-5.
T. Gopalakrishnan, S. Ramakrishnan, K. Balasamy and A. S. Muthananda Murugavel, "Semi fragile watermarking using Gaussian mixture model for malicious image attacks," 2011 World Congress on Information and Communication Technologies, Mumbai, India, 2011, pp. 120-125, https://doi.org/10.1109/WICT.2011.6141229.
Renuka Devi, K., Suganyadevi, S and Balasamy, K. “Healthcare Data Analysis U sing Deep Learning Paradigm”. Deep Learning for Cognitive Computing Systems: Technological Advancements and Applications, edited by M.G. Sumithra, Rajesh Kumar Dhanaraj, Celestine Iwendi and Anto Merline Manoharan, Berlin, Boston:De Gruyter, 2023, pp. 129–148. https://doi.org/10.1515/9783110750584-008.
Shamia, D., Balasamy, K., and Suganyadevi, S. “A secure framework for medical image by integrating watermarking and encryption through fuzzy based roi selection”, Journal of Intelligent & Fuzzy systems, 2023, Vol. 44, no.5, pp.7449-7457, https://doi.org/10.3233/JIFS-222618.
E. Elyan, P. Vuttipittayamongkol, P. Johnston, K. Martin, K. McPherson, C.F. Moreno-García, C. Jayne, M.M.K. Sarker, Computer vision and machine learning for medical image analysis: recent advances, challenges, and way forward, Artif. Intell. Surg. 2 (1) (2022) 24–45. https://doi.org/10.20517/ais.2021.15
O. Ayo-Farai, B.A. Olaide, C.P. Maduka, C.C. Okongwu, Engineering innovations in healthcare: a review of developments in the USA, Eng. Sci. Technol. J. 4 (6) (2023) 381–400. https://doi.org/10.51594/estj.v4i6.638
S. Yin, C. Fu, S. Zhao, K. Li, X. Sun, T. Xu, and E. Chen, “A survey on multimodal large language models,” arXiv preprint arXiv:2306.13549, 2023. https://doi.org/10.1093/nsr/nwae403
D. Hao, M. Meng, Y. Gao, X. Lou and W. Kong, "Step-Wise Prompting Meets Uncertainty-Aware Dynamic Fusion for Robust EEG-Visual Emotion Recognition," in IEEE Transactions on Affective Computing, vol. 17, no. 1, pp. 694-707, Jan.-March 2026, https://doi.org/ 10.1109/TAFFC.2025.3632304
Z. Lin, X. Hu, Y. Zhang, Z. Chen, Z. Fang, X. Chen, A. Li, P. Vepakomma, and Y. Gao, “SplitLoRA: A Split Parameter-Efficient Fine-Tuning Framework for Large Language Models,” arXiv preprint arXiv:2407.00952, 2024. https://doi.org/10.48550/arXiv.2407.00952
S. Hu, Z. Fang, Z. Fang, Y. Deng, X. Chen, Y. Fang, and S. T. W. Kwong, “Agentscomerge: Large language model empowered collaborative decision making for ramp merging,” IEEE Transactions on Mobile Computing, 2025. https://doi.org/10.1109/TMC.2025.3564163
S. Hu, Z. Fang, Z. Fang, Y. Deng, X. Chen, and Y. Fang, “Agentscodriver: Large language model empowered collaborative driving with lifelong learning,” arXiv preprint arXiv:2404.06345, 2024.
https://doi.org/10.48550/arXiv.2404.06345
K. Wu, B. Jiang, Z. Jiang, Q. He, D. Luo, S. Wang, Q. Liu, and C. Wang, “Noiseboost: Alleviating hallucination with noise perturbation for multimodal large language models,” arXiv preprint arXiv:2405.20081, 2024. https://doi.org/10.48550/arXiv.2405.20081
Lin, Qinghua, Guang-Hai Liu, Zuoyong Li, Yang Li, Yuting Jiang, and Xiang Wu. "Multimodal Medical Image Classification via Synergistic Learning Pre-training." arXiv preprint arXiv:2509.17492 (2025). https://doi.org/10.48550/arXiv.2509.17492
Q. Wang, L. Zhan, P. Thompson, and J. Zhou, “Multimodal learning with incomplete modalities by knowledge distillation,” in Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, August 2020, pp. 1828–1838. https://doi.org/10.1145/3394486.3403234
J. Zhao, R. Li, and Q. Jin, “Missing modality imagination network for emotion recognition with uncertain missing modalities,” in Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers), 2021, pp. 2608–2618. https://doi.org/10.18653/v1/2021.acl-long.203
K. Zhou, J. Li, Y. Xiao, J. Yang, J. Cheng, W. Liu, W. Luo, J. Liu, and S. Gao, “Memorizing structure-texture correspondence for image anomaly detection,” IEEE Transactions on Neural Networks and Learning Systems, vol. 33, no. 6, pp. 2335–2349, 2021. https://doi.org/10.1109/TNNLS.2021.3101403
Copyright (c) 2026 Krishnakumar B, Thanga Parvathi, K. Nithya, M. Pyingkodi, Kunchanapalli Rama Krishna, Jeevitha R
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution-ShareAlikel 4.0 International (CC BY-SA 4.0) 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) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
.png)
.png)
.png)
.png)
.png)
