Journal of Soft Computing and Information Technology

Journal of Soft Computing and Information Technology

Automatic Segmentation of Multiple Sclerosis Lesions on MRI Images Using Adaptive Residual U-Net

Document Type : Persian Original Article

Authors
Ali Fahmi Jafargholkhanloo 1
Mousa Shamsi 2
Mahdi Bashiri Bawil 2
1 Department of Engineering Sciences, Faculty of Advanced Technologies, University of Mohaghegh Ardabili, Namin, Iran.
2 Department of Bioelectric, Faculty of Biomedical Engineering, Sahand University of Technology, Tabriz, Iran.
10.22034/jscit.2026.550141.2177
Abstract
Multiple Sclerosis (MS) is a chronic and neurodegenerative disorder of the central nervous system, whose diagnosis, classification, and monitoring are primarily performed through Magnetic Resonance Imaging. Accurate segmentation of MS lesion boundaries plays an essential role in early diagnosis, disease severity assessment, and treatment monitoring. However, despite the reliability of manual segmentation, this process is labor-intensive, time-consuming, and requires substantial expertise. Current automated methods show limited sensitivity to small lesions and poor generalization across different acquisition parameters. A residual U-Net architecture incorporating adaptive patch extraction and hybrid loss optimization was developed. The approach combines probabilistic sampling to address class imbalance, residual connections to enhance gradient flow, and a composite Dice-Focal loss function tailored for small object segmentation. The proposed architecture was evaluated on the MSLesSeg dataset comprising 115 FLAIR sequences from 75 patients across multiple imaging centers. Experimental results demonstrate that the proposed architecture has an LTPR of 79.92%, a precision of 72.77%, a recall of 72.22%, and a Dice score of 72.49%, outperforming the classical U-Net and other compared methods in accurate and sensitive detection of MS lesions. Integration of adaptive sampling with residual U-Net architecture effectively addresses fundamental challenges in MS lesion segmentation.
Keywords
Adaptive Patch Extraction
Deep Learning
Multiple Sclerosis Lesions
Medical Image Segmentation
Residual Network
Neuroimaging
[1] F. Guarnera, A. Rondinella, E. Crispino, G. Russo, C. Di Lorenzo, D. Maimone, ... and S. Battiato, “MSLesSeg: baseline and benchmarking of a new Multiple Sclerosis Lesion Segmentation dataset,” Scientific Data, vol. 12, no. 920, 2025. doi: https://doi.org/10.1038/s41597-025-05250-y.
[2] M. Popa, G. A. Vișa, and C. R. Șofariu, “PediMS: A Pediatric Multiple Sclerosis Lesion Segmentation Dataset,” Scientific Data, vol. 12, no. 1184, 2025. doi: https://doi.org/10.1038/s41597-025-05346-5.
[3] S. Valverde, M. Cabezas, E. Roura, S. González-Villà, D. Pareto, J. C. Vilanova, ... and X. Lladó, “Improving automated multiple sclerosis lesion segmentation with a cascaded 3D convolutional neural network approach,” NeuroImage, vol. 155, pp. 159-168, 2017. doi: https://doi.org/10.1016/j.neuroimage.2017.04.034.
[4] S. Roy, J. A. Butman, D. S. Reich, P. A. Calabresi and D. L. Pham, “Multiple sclerosis lesion segmentation from brain MRI via fully convolutional neural networks,” arXiv preprint arXiv:1803.09172, 2018. doi: https://doi.org/10.48550/arXiv.1803.09172.
[5] S. Aslani, M. Dayan, L. Storelli, M. Filippi, V. Murino, M. A. Rocca and D. Sona, “Multi-branch convolutional neural network for multiple sclerosis lesion segmentation,” NeuroImage, 196, 1-15, 2019. doi: https://doi.org/10.1016/j.neuroimage.2019.03.068.
[6] W. Wang, C. Chen, M. Ding, H. Yu, S. Zha and J. Li, “Transbts: Multimodal brain tumor segmentation using transformer,” In International conference on medical image computing and computer-assisted intervention, pp. 109-119, 2021. doi: https://doi.org/10.1007/978-3-030-87193-2_11.
[7] A. Hatamizadeh, Y. Tang, V. Nath, D. Yang, A. Myronenko, B. Landman, ... and D. Xu, “Unetr: Transformers for 3d medical image segmentation,” In Proceedings of the IEEE/CVF winter conference on applications of computer vision, pp. 574-584, 2022. doi: https://doi.org/10.1109/WACV51458.2022.00181.
[8] A. Hatamizadeh, V. Nath, Y. Tang, D. Yang, H. R. Roth, and D. Xu, “Swin unetr: Swin transformers for semantic segmentation of brain tumors in mri images,” In International MICCAI brainlesion workshop, pp. 272-284, 2021. doi: https://doi.org/10.1007/978-3-031-08999-2_22.
[9] H. Lu, Y. She, J. Tie and S. Xu, “Half-UNet: A simplified U-Net architecture for medical image segmentation,” Frontiers in neuroinformatics, vol. 16, 2022. doi: https://doi.org/10.3389/fninf.2022.911679.
[10] M. Sadeghibakhi, H. Pourreza, and H. Mahyar, “Multiple sclerosis lesions segmentation using attention-based CNNs in FLAIR images,” IEEE journal of translational engineering in health and medicine, vol. 10, pp. 1-11, 2022. doi: https://doi.org/10.1109/JTEHM.2022.3172025.
[11] A. Rondinella, E. Crispino, F. Guarnera, O. Giudice, A. Ortis, G. Russo, ... and S. Battiato, “Boosting multiple sclerosis lesion segmentation through attention mechanism. Computers in Biology and Medicine, vol. 161, pp. 107021, 2023. doi: https://doi.org/10.1016/j.compbiomed.2023.107021.
[12] A. Rondinella, F. Guarnera, O. Giudice, A. Ortis, G. Russo, E. Crispino, ... and S. Battiato, “Enhancing multiple sclerosis lesion segmentation in multimodal mri scans with diffusion models,” In 2023 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), pp. 3733-3740, 2023. doi: https://doi.org/10.1109/BIBM58861.2023.10385334.
[13] A. P. De Rosa, M. Benedetto, S. Tagliaferri, F. Bardozzo, A. D’Ambrosio, A. Bisecco, ... and F. Esposito, “Consensus of algorithms for lesion segmentation in brain MRI studies of multiple sclerosis,” Scientific Reports, vol. 14, no. 1, 2024. doi: https://doi.org/10.1038/s41598-024-72649-9.
‏[14] S. R. Hashemi, S. S. M. Salehi, D. Erdogmus, S. P. Prabhu, S. K. Warfield, and A. Gholipour, “Asymmetric loss functions and deep densely-connected networks for highly-imbalanced medical image segmentation: Application to multiple sclerosis lesion detection,” IEEE Access, vol. 7, pp. 1721-1735, 2018. doi: https://doi.org/10.1109/access.2018.2886371.
[15] E. Essa, D. Aldesouky, S. E. Hussein and M. Z. Rashad, “Neuro-fuzzy patch-wise R-CNN for multiple sclerosis segmentation,” Medical & Biological Engineering & Computing, vol. 58, no. 9, pp. 2161-2175, 2020. doi: https://doi.org/10.1007/s11517-020-02225-6.
[16] O. Ronneberger, P. Fischer and T. Brox, “U-net: Convolutional networks for biomedical image segmentation,” In International Conference on Medical image computing and computer-assisted intervention, pp. 234-241, 2015. doi: https://doi.org/10.48550/arXiv.1505.04597.
[17] K. He, X. Zhang, S. Ren and J. Sun, “Deep residual learning for image recognition,” In Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 770-778, 2016. doi: https://doi.org/10.1109/CVPR.2016.90.
[18] N. Abraham and N.M. Khan, “A novel focal tversky loss function with improved attention u-net for lesion segmentation,” In 2019 IEEE 16th international symposium on biomedical imaging, pp. 683-687, 2019. doi: https://doi.org/10.1109/ISBI.2019.8759329.
[19] T. Y. Lin, P. Goyal, R. Girshick, K. He and P. Dollár, “Focal loss for dense object detection. In Proceedings of the IEEE international conference on computer vision, pp. 2980-2988, 2017. doi: https://doi.org/10.1109/ICCV.2017.324.
 
 
Journal of Soft Computing and Information Technology
Volume 15, Issue 2
Summer - August
Summer 2026
Pages 1-15