This paper explores the application of the U-Net architecture for intracranial hemorrhage segmentation, with a focus on enhancing segmentation accuracy through the incorporation of texture enhancement techniques based on the Riesz fractional order derivatives. The study begins by conducting a review of related works in the field of computed tomography (CT) scan segmentation. At this stage also a suitable dataset is selected. Initially it is used to train the UNet, one of the widely adopted deep learning models in the field of medical image segmentation. Training is performed using parallel algorithm based on CUDA technology. The obtained results are compared with the established baseline for this dataset, assessing segmentation accuracy using the Jaccard and Dice coefficients. Subsequently, the study investigates a texture enhancement technique based on the Riesz fractional order derivatives, applied to the CT-scan images from the dataset. This technique aims to capture finer details and subtle textures that may contribute to improved segmentation accuracy. The U-Net model is then retrained and validated on the texture-enhanced images, and the experimental results are analyzed. The study reveals a modest yet notable enhancement in accuracy, as measured by the Jaccard and Dice coefficients, demonstrating the potential of the proposed texture enhancement technique in refining intracranial hemorrhage segmentation.
[1] S. Chilamkurthy, R. Ghosh, S. Tanamala, M. Biviji, N. G. Campeau, V. K. Venugopal, V. Mahajan, P. Rao, P. Warier, Deep learning algorithms for detection of critical findings in head CT scans: a retrospective study. Lancet 01.12 (2018): 2388-2396. https://doi.org/10.1016/S0140-6736(18)31645-3
[2] Radiological Society of North America, RSNA Intracranial Hemorrhage Detection, 2019. URL: https://www.kaggle.com/c/rsna-intracranial-hemorrhage-detection/overview.
[3] M. D. Hssayeni, M.S. Croock, A. D. Salman, H. F. Al-khafaji, Z. A. Yahya, B. Ghoraani, Intracranial Hemorrhage Segmentation Using a Deep Convolutional Model, Data 5 (2020): 14-32. https://doi.org/10.3390/data5010014
[4] M. D. Hssayeni, Computed Tomography Images for Intracranial Hemorrhage Detection and Segmentation, version 1.3.1, PhysioNet, 2020. doi: 10.13026/4nae-zg36.
[5] A. Goldberger, L. Amaral, L. Glass, J. Hausdorff, P. C. Ivanov, R. Mark, J. E. Mietus, G. B. Moody, C. K. Peng, H. E. Stanley, PhysioBank, PhysioToolkit, and PhysioNet: Components of a new research resource for complex physiologic signals, 2000. URL: https://physionet.org. https://doi.org/10.1161/01.CIR.101.23.e215
[6] O. Ronneberger, P. Fischer, T. Brox, U-net: convolutional networks for biomedical image segmentation, in: Proceedings of the International Conference on Medical image computing and computer-assisted intervention, Springer, Cham, 2015, pp. 234-241. doi: 10.1007/978-3-319-24574-4_28.
[7] Q. Yu, F. Liu, I. Turner, K. Burrage, V. Vegh, The use of a Riesz fractional differential-based approach for texture enhancement in image processing, ANZIAM Journal 54 (2012): 590-607. https://doi.org/10.1007/978-3-319-24574-4_28
[8] J. Hamid, I. Rabha, Texture Enhancement for Medical Images Based on Fractional Differential Masks, Discrete Dynamics in Nature and Society, 28.03 (2013). https://doi.org/10.1155/2013/618536
[9] Y.-F. Pu, J.-L. Zhou, X. Yuan, Fractional Differential Mask: A Fractional Differential-Based Approach for Multiscale Texture Enhancement, IEEE Transactions on Image Processing 19 (2010): 491-511. https://doi.org/10.1109/TIP.2009.2035980
[10] Ya. Sokolovskyy, M. Levkovych and I. Sokolovskyy, The study of heat transfer and stress-strain state of a material, taking into account its fractal structure. Mathematical Modeling and Computing. 7(2), 2020, pp. 400–409. https://doi.org/10.23939/mmc2020.02.400
[11] Ya. Sokolovskyy, M. Levkovych, O. Mokrytska, and Ya. Kaplunskyy, Mathematical models of biophysical processes taking into account memory effects and self-similarity, CEUR Workshop Proceedings, 2018, vol. 2255, pp. 215–228.
[12] F. Liu, P. Zhuang, V. Anh, I. Turner, and K. Burrage, Stability and convergence of the difference methods for the space–time fractional advection–diffusion equation, Applied Mathematics and Computation, vol. 191, issue 1 (2007): 12-20. https://doi.org/10.1016/j.amc.2006.08.162
[13] M. Nadrljanski, A. Campos, R. Chieng, et al. Computed tomography. Reference article, Radiopaedia.org, 2024. https://doi.org/10.53347/rID-9027
[14] K. Greenway, R. Sharma, D. Vargas Carvajal, et al. Hounsfield unit. Reference article, Radiopaedia.org, 2024. https://doi.org/10.53347/rID-38181
[15] A. Paszke, S. Gross, S. Chintala, G. Chanan, E. Yang, Z. DeVito, Z. Lin, A. Desmaison, L. Antiga, A. Lerer, Automatic differentiation in PyTorch, in: Proceedings of the 31st Conference on Neural Information Processing Systems, NIPS 2017, Long Beach, CA, USA, 2017. URL: https://openreview.net/forum?id=BJJsrmfCZ.
[16] PyTorch Contributors, PyTorch Documentation, 2023. URL: https://pytorch.org/docs/stable/index.html.
[17] Christian S. Perone, cclauss, Elvis Saravia, Pedro Lemos Ballesterі, MohitTare, “perone/medicaltorch: Release v0.2”, Zenodo, 2018. doi: 10.5281/zenodo.1495335.
[18] Colaboratory, Frequently Asked Questions, 2023. URL: https://research.google.com/colaboratory/faq.html.
[19] NVIDIA Corporation & Affiliates, CUDA Toolkit Documentation 12.2, 2023. URL: https://docs.nvidia.com/cuda/archive/12.2.0/.
[20] L. G. Shapiro, G. C. Stockman, Co-occurrence Matrices and Features, in: Computer Vision, 1st. ed., Pearson, 2001, pp. 240-243.