Method for registering multiphase CT images with an adaptive weight function

The article proposes an improved method for recording multiphase CT scans of the liver, based on a modified displacement formula that combines global affinity transformation, a local B-spline deformation model, adaptive correction based on an intensity gradient and a noise component to increase robustness. The key element is the weight function E(x), which takes into account local differences in tissue density, it limits deformations in dense structures and enhances them in soft tissues and pathologies. The algorithm is implemented in two consecutive stages – first affine, then B-spline registration – using an extended liver mask and deferred cropping of images, which significantly improves convergence and accuracy. Experiments on a set of clinical cases have shown the superiority of the proposed approach over standard methods. A particularly significant increase was achieved for small pathologies (1–1000 voxels), the average DICE coefficient increased from 0.5737 (affinity registration) to 0.6277. The method demonstrates high resistance to artifacts caused by the patient's breathing, noise and contrast inhomogeneity, and also ensures accurate alignment at the boundaries of objects. The results confirm the clinical applicability of the approach for diagnosis, analysis of liver disease dynamics and treatment planning.

1. Zitová B., Flusser J. Image Registration Methods: A Survey. Image and Vision Computing. 2003;21(11):977–1000. https://doi.org/10.1016/S0262-8856(03)00137-9

2. Rueckert D., Sonoda L.I., Hayes C., Hill D.L.G., Leach M.O., Hawkes D.J. Nonrigid Registration Using Free-Form Deformations: Application to Breast MR Images. IEEE Transactions on Medical Imaging. 1999;18(8):712–721. https://doi.org/10.1109/42.796284

3. Avants B.B., Epstein C.L., Grossman M., Gee J.C. Symmetric Diffeomorphic Image Registration with Cross-Correlation: Evaluating Automated Labeling of Elderly and Neurodegenerative Brain. Medical Image Analysis. 2008;12(1):26–41. https://doi.org/10.1016/j.media.2007.06.004

4. Vercauteren T., Pennec X., Perchant A., Ayache N. Symmetric Log-Domain Diffeomorphic Registration: A Demons-Based Approach. In: Medical Image Computing and Computer-Assisted Intervention – MICCAI 2008: 11th International Conference: Proceedings, Part I, 06–10 September 2008, New York, USA. Berlin, Heidelberg: Springer; 2008. P. 754–761. https://doi.org/10.1007/978-3-540-85988-8_90

5. Fischer B., Modersitzki J. Fast Diffusion Registration. AMS Contemporary Mathematics, Inverse Problems, Image Analysis, and Medical Imaging. 2002;313:117–127. https://doi.org/10.1090/conm/313/05372

6. Balakrishnan G., Zhao A., Sabuncu M.R., Guttag J., Dalca A.V. VoxelMorph: A Learning Framework for Deformable Medical Image Registration. IEEE Transactions on Medical Imaging. 2019;38(8):1788–1800. https://doi.org/10.1109/TMI.2019.2897538

7. Hatamizadeh A., Tang Yu., Nath V., et al. UNETR: Transformers for 3D Medical Image Segmentation. In: 2022 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 03–08 January 2022, Waikoloa, HI, USA. IEEE; 2022. P. 1748–1758. https://doi.org/10.1109/WACV51458.2022.00181

8. Guo Zh., Li X., Huang H., Guo N., Li Q. Medical Image Segmentation Based on Multi-Modal Convolutional Neural Network: Study on Image Fusion Schemes. In: 2018 IEEE 15th International Symposium on Biomedical Imaging (ISBI 2018), 04–07 April 2018, Washington, DC, USA. IEEE; 2018. P. 903–907. https://doi.org/10.1109/ISBI.2018.8363717

9. Kulikov A.A., Kashirina I.L., Savkina E.F. Segmentation of Liver Volumetric Lesions in Multiphase CT Images Using the nnU-Net Framework. Modeling, Optimization and Information Technology. 2025;13(1). (In Russ.). https://doi.org/10.26102/2310-6018/2025.48.1.040

10. Salehi S.S.M., Erdogmus D., Gholipour A. Tversky Loss Function for Image Segmentation Using 3D Fully Convolutional Deep Networks. In: Machine Learning in Medical Imaging: 8th International Workshop – MLMI 2017, Held in Conjunction with MICCAI 2017, 10 September 2017, Quebec City, QC, Canada. Cham: Springer; 2017. P. 379–387. https://doi.org/10.1007/978-3-319-67389-9_44

11. Khoong W.H. BUSU-Net: An Ensemble U-Net Framework for Medical Image Segmentation. arXiv. URL: https://arxiv.org/abs/2003.01581 [Accessed 10th September 2025].