CNN–TRANSFORMER HYBRID ARCHITECTURE AND SELF-SUPERVISED LEARNING APPROACHES FOR BRAIN TUMOR SEGMENTATION BASED ON 3D MRI IMAGES
Abdukadirova G. K. & Abduqodirova M. I.
Annotatsiya
Brain tumor segmentation from three-dimensional Magnetic Resonance Imaging (3D MRI) is one of the most challenging and clinically critical tasks in medical image analysis. Conventional convolutional neural networks (CNNs), while effective for local feature extraction, are inherently limited in capturing the long-range spatial dependencies essential for precise volumetric segmentation. Recent advances in Vision Transformers and self-supervised learning (SSL) paradigms offer transformative potential for overcoming these limitations, particularly in data-scarce clinical settings. This article presents a comprehensive review and technical analysis of hybrid CNN–Transformer architectures and SSL-based pretraining strategies applied to 3D brain tumor segmentation. The study examines state-of-the-art models including TransUNet, SwinUNETR, and nnFormer, evaluates SSL pretext tasks such as Masked Volume Modeling and contrastive learning, and critically analyses performance on the BraTS 2021 benchmark dataset. Experimental comparisons demonstrate that SwinUNETR with SSL pretraining achieves a mean whole-tumour Dice Similarity Coefficient (DSC) of 91.3%, significantly outperforming purely convolutional baselines. The article further discusses open challenges, computational trade-offs, and directions for future research in clinically deployable AI-driven neuro-oncology diagnostics.
Adabiyotlar
- Ostrom Q.T., Cioffi G., Gittleman H. et al. CBTRUS Statistical Report: primary brain and other central nervous system tumors diagnosed in the United States in 2014–2018 // Neuro-Oncology. — 2021. — Vol. 23, No. 12, Suppl. 2. — P. iii1–iii105. DOI: 10.1093/neuonc/noab200
- Dosovitskiy A., Beyer L., Kolesnikov A. et al. An image is worth 16x16 words: transformers for image recognition at scale // International Conference on Learning Representations (ICLR). — 2021. arXiv: 2010.11929
- Chen J., Lu Y., Yu Q. et al. TransUNet: transformers make strong encoders for medical image segmentation // arXiv preprint. — 2021. arXiv: 2102.04306
- Tang Y., Yang D., Li W. et al. Self-supervised pre-training of Swin transformers for 3D medical image analysis // Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). — 2022. — P. 20730–20740. DOI: 10.1109/CVPR52688.2022.02007
- Zhou H.-Y., Guo J., Zhang Y. et al. nnFormer: interleaved transformer for volumetric segmentation // arXiv preprint. — 2021. arXiv: 2109.03201
- Baid U., Ghodasara S., Mohan S. et al. The RSNA-ASNR-MICCAI BraTS 2021 benchmark on brain tumor segmentation and radiogenomic classification // arXiv preprint. — 2021. arXiv: 2107.02314
- Isensee F., Jaeger P.F., Kohl S.A.A. et al. nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation // Nature Methods. — 2021. — Vol. 18, No. 2. — P. 203–211. DOI: 10.1038/s41592-020-01008-z
- Liu Z., Lin Y., Cao Y. et al. Swin transformer: hierarchical vision transformer using shifted windows // Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV). — 2021. — P. 10012–10022. DOI: 10.1109/ICCV48922.2021.00986
- Lee C.-Y., Xie S., Gallagher P. et al. Deeply-supervised nets // Artificial Intelligence and Statistics (AISTATS). — PMLR. — 2015. — Vol. 38. — P. 562–570.
- Hatamizadeh A., Nath V., Tang Y. et al. Swin UNETR: Swin transformers for semantic segmentation of brain tumors in MRI images // International MICCAI Brainlesion Workshop. — Springer, Cham. — 2022. — P. 272–284. DOI: 10.1007/978-3-031-08999-2_22
- Liu X., Zhang F., Hou Z. et al. Self-supervised learning: generative or contrastive // IEEE Transactions on Knowledge and Data Engineering. — 2023. — Vol. 35, No. 1. — P. 857–876. DOI: 10.1109/TKDE.2021.3090866
- Devlin J., Chang M.-W., Lee K. et al. BERT: pre-training of deep bidirectional transformers for language understanding // Proceedings of NAACL-HLT. — 2019. — P. 4171–4186. DOI: 10.18653/v1/N19-1423
- He K., Chen X., Xie S. et al. Masked autoencoders are scalable vision learners // Proceedings of the IEEE/CVF CVPR. — 2022. — P. 16000–16009. DOI: 10.1109/CVPR52688.2022.01553
- Chen T., Kornblith S., Norouzi M., Hinton G. A simple framework for contrastive learning of visual representations // Proceedings of the 37th International Conference on Machine Learning (ICML). — PMLR. — 2020. — Vol. 119. — P. 1597–1607.
- Zhu J., Li Y., Hu Y. et al. Rubik's cube+: a self-supervised feature learning framework for 3D medical image analysis // Medical Image Analysis. — 2020. — Vol. 64. — 101746. DOI: 10.1016/j.media.2020.101746
- Cardoso M.J., Li W., Brown R. et al. MONAI: an open-source framework for deep learning in healthcare // arXiv preprint. — 2022. arXiv: 2211.02701
- Isensee F., Jaeger P.F., Kohl S.A.A. et al. nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation // Nature Methods. — 2021. — Vol. 18. — P. 203–211.
- 23. Guan H., Liu M. Domain adaptation for medical image analysis: a survey // IEEE Transactions on Biomedical Engineering. — 2022. — Vol. 69, No. 3. — P. 1173–1185. DOI: 10.1109/TBME.2021.3117407
- Selvaraju R.R., Cogswell M., Das A. et al. Grad-CAM: visual explanations from deep networks via gradient-based localization // International Journal of Computer Vision. — 2020. — Vol. 128, No. 2. — P. 336–359. DOI: 10.1007/s11263-019-01228-7