隨著臨床醫學影像技術的發展，CT、X-Ray、MRI等都是常見的醫學影像儀器，但存在輻射危害、測量時間長以及成本高昂等限制。而電阻抗斷層成像（Electrical Impedance Tomography, EIT）因具備非侵入性、低成本和可即時監測等優點，在臨床應用面上具有相當大的潛力。然而，EIT系統的主要缺點在於空間解析度低與影像失真嚴重，導致影像可辨識度不佳，限制了其在臨床上的應用。因此，提升EIT影像品質與改善重建準確性，成為目前研究的重要課題。目前研究多應用於監測肺部氣體變化，透過分析氣體在肺部進出時的導電率變化，並根據此辨識影像中肺部區域。然而，針對大腿影像卻沒有能辨識肌肉區域方法，現階段仍缺乏能夠有效並準確區分肌肉組織分布的方法。本研究針對此問題，提出一套結合自適應八電極EIT系統與機器學習校正模型之影像校正方法。首先，利用所建立的 EIT 硬體系統進行大腿電導率測量，接著從重建影像中擷取多項統計與幾何特徵，並以極限梯度提升演算法(Extreme Gradient Boosting, XGboost)進行校正，生成與實際解剖結構相符的校正影像。實驗結果顯示，所提出方法能有效提升影像辨識能力，使肌肉區域的導電率分布更加接近生理實際情況，並顯著改善傳統重建影像中常見的邊界模糊與數值誤差問題。未來有望應用於肌少症患者之肌肉變化監測與復健評估，提供即時且量化的影像輔助資訊，具有降低人力成本與提升臨床判斷效率的潛在價值。

With the advancement of clinical medical imaging technologies, instruments such as computed tomography (CT), X-ray, and magnetic resonance imaging (MRI) have become common diagnostic tools. However, these imaging modalities still face limitations, including radiation exposure, long acquisition time, and high operational cost. Electrical Impedance Tomography (EIT), on the other hand, possesses several advantages such as non-invasiveness, low cost, and real-time monitoring capability, making it a promising alternative for clinical applications. Nevertheless, the main drawback of EIT lies in its low spatial resolution and significant image distortion, which limit its accuracy and diagnostic usability. Therefore, improving image quality and reconstruction accuracy has become a key research focus in recent years. Most current studies have applied EIT to lung ventilation monitoring, where conductivity variations caused by airflow are analyzed to distinguish lung regions in the reconstructed images. However, for thigh muscle imaging, there remains a lack of effective methods capable of accurately identifying muscle regions and differentiating tissue distribution. To address this issue, this study proposes an image correction approach that integrates an eight-electrode EIT system with a machine learning–based correction model. The developed EIT hardware system measures the conductivity distribution of the thigh, from which multiple statistical and geometric features are extracted. These features are then utilized in an Extreme Gradient Boosting (XGBoost) regression model to generate corrected images that better correspond to the actual anatomical structures.Experimental results demonstrate that the proposed method effectively enhances image interpretability, making the conductivity distribution within muscle regions more consistent with physiological conditions. Furthermore, it significantly improves the boundary clarity and reduces numerical errors commonly found in conventional EIT reconstructions. In the future, this system is expected to be applied to the monitoring of muscle alterations and rehabilitation assessment in patients with sarcopenia, offering real-time and quantitative imaging support for clinical evaluation. Such an approach holds significant potential for reducing manpower requirements and improving the efficiency of clinical decision-making.

[1] F. Deconinck, “Nuclear imaging in the realm of medical imaging,” Nuclear Instruments and Methods in Physic Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, vol. 509, no. 1, pp. 213–228, Aug, 2003.
[2] D. C. Barber and B. H. Brown, “Recent developments in applied potential tomography APT,” Information Processing in Medical Imaging. Springer, pp.106–107, Jan, 1986.
[3] T. Piraino, “An Introduction to the Clinical Application and Interpretation of Electrical Impedance Tomography,” Respiratory Care, vol. 67, no. 6, pp. 721–729, Jun, 2022.
[4] S. Leonhardt and B. Lachmann, “Electrical impedance tomography: the holy grail of ventilation and perfusion monitoring ?,” Intensive Care Medicine, vol. 38, no. 12, pp. 1917–1929, Dec, 2012.
[5] S. Grimnes and O. G. Martinsen, “Bioimpedance and Bioelectricity Basics (Biomedical Engineering),” Academic Press, pp. 93-137, May, 2008.
[6] R. J. Sadleir and R. A. Fox, “Detection and quantification of intraperitoneal fluid using electrical impedance tomography,” IEEE Transactions on Biomedical Engineering, vol. 48, no. 4, pp. 484–491, Apr, 2001.
[7] Y. Shi, X. Bai, J. Yang, X. Wu, and L. Wang, “Optimized measurement methods and systems for the dielectric properties of active biological tissues in the 10Hz-100 MHz frequency range,” Frontiers in Physiology, vol. 16, pp. 1-11, Jan, 2025.
[8] M. Mirhoseini, Z. R. Gatabi, S. Das, S. Joveini, and I. R. Gatabi, “Applications of Electrical Impedance Tomography in Neurology,” Basic and Clinical Neuroscience, vol. 13, no. 5, pp. 595–608, Sep, 2022.
[9] M. Damijan, P. Natasa and H. Francis, “Electric Properties of Tissues,” Wiley Encyclopedia of Biomedical Engineering, vol. 6, pp. 1-13, Apr, 2006.
[10] F. Pennati et al., “Electrical Impedance Tomography: From the Traditional Design to the Novel Frontier of Wearables,” Sensors, vol. 23, no. 3, pp.1182, Jan, 2023.
[11] A. Nissinen, V. P. Kolehmainen and J. P. Kaipio, “Compensation of Modelling Errors Due to Unknown Domain Boundary in Electrical Impedance Tomography,” IEEE Transactions on Medical Imaging, vol. 30, no. 2, pp. 231–242, Feb, 2011.
[12] G. Xu et al., “3-D electrical impedance tomography forward problem with finite element method,” IEEE Transactions on Magnetics, vol. 41, no. 5, pp. 1832–1835, May, 2005.
[13] D. Todor, G. Copeland and L. Cheran, “Electrical impedance tomography: An improved model for solving the forward problem,” Physica Medica, vol. 13, pp. 358–361, Dec, 1997.
[14] K. Y. Kim, B. S. Kim, M. C. Kim and S. Kim, “Dynamic inverse obstacle problems with electrical impedance tomography,” Mathematics and Computers in Simulation, vol. 66, no. 4, pp. 399–408, Jul, 2004.
[15] D. Sbarbaro, M. Vauhkonen and T. A. Johansen, “State estimation and inverse problems in electrical impedance tomography: observability, convergence and regularization,” Inverse Problems, vol. 31, no. 4, pp. 1-28, Apr, 2015.
[16] D. Maciejewski, Z. Putowski, M. Czok and L. J. Krzych, “Electrical impedance tomography as a tool for monitoring mechanical ventilation. An introduction to the technique,” Advances in Medical Sciences, vol. 66, no. 2, pp. 388–395, Sep, 2021.
[17] H. Zhang, R. Cai, Y. Yang and N. Li, “Unsupervised Deep-Learning-Based Error Image Prior (DLEIP) Algorithm for Lung Electrical Impedance Tomography (EIT),” IEEE Internet of Things Journal, vol. 12, no. 15, pp. 31399–31407, Aug, 2025.
[18] S. Zhu, K. Li, S. Yue and L. Liu, “Clustering-Based Reconstruction Algorithm for Electrical Impedance Tomography,” IEEE Transactions on Instrumentation and Measurement, vol. 73, pp. 1–11, Aug, 2024.
[19] Z. Liu, Z. Chen, H. Fang, Y. Feng and Y. Yang, “Regularized Shallow Image Prior for Electrical Impedance Tomography,” IEEE Transactions on Instrumentation and Measurement, vol. 74, pp. 1-11, Feb, 2025.
[20] X. Li et al., “An image reconstruction framework based on deep neural network for electrical impedance tomography,” IEEE International Conference on Image Processing (ICIP), pp. 3585–3589, Sep, 2017.
[21] T. Zhang et al., “Advances of deep learning in electrical impedance tomography image reconstruction,” Front. Bioeng. Biotechno, vol. 10, pp.1-17, Dec, 2022.
[22] T. Li, T. J. Kao, D. Isaacson, J. C. Newell and G. J. Saulnier, “Adaptive Kaczmarz method for image reconstruction in electrical impedance tomography,” Physiological Measurement, vol. 34, no. 6, pp. 595, May, 2013.
[23] F. Li, B. Chen, Y. Wu, K. Liu, T. Zhou and J. F. Yao, “Research on a COPD Diagnosis Method Based on Electrical Impedance Tomography Imaging,” Progress in Biochemistry and Biophysics, vol. 52, no. 7, pp. 1866–1877, Jul, 2025.
[24] M. Bodenstein, M. David and K. Markstaller, “Principles of electrical impedance tomography and its clinical application,” Critical Care Medicine, vol. 37, no. 2, pp. 713–724, Feb, 2009.
[25] Y. Li, L. Rao, R. He et al., “A novel combination method of electrical impedance-tomography inverse problem for brain imaging,” IEEE Transactions on Magnetics, vol. 41, no. 5, pp. 1848–1851, May, 2005.
[26] Y. Guo, L. Zhu, M. Wang, Y. Jiang, M. Soleimani and M. Zhang, “Capacitively Coupled Electrical Impedance Tomography in Lung Imaging,” IEEE Sensors Journal, vol. 24, no. 20, pp. 33072–33082, Oct, 2024.
[27] Z. Cui, X. Liu, H. Qu and H. Wang, “Technical Principles and Clinical Applications of Electrical Impedance Tomography in Pulmonary Monitoring,” Sensors, vol. 24, no. 14, pp. 4539, Jul, 2024.
[28] P. A. Sejati, B. Sun, P. N. Darma, T. Shirai, K. Narita and M. Takei, “Multinode Electrical Impedance Tomography (mnEIT) Throughout Whole-Body Electrical Muscle Stimulation (wbEMS),” IEEE Transactions on Instrumentation and Measurement, vol. 72, pp. 1–14, Jun, 2023.
[29] B. Sun, P. A. Sejati, T. Shirai and M. Takei, “Long-term phase angle muscle imaging under electrical muscle stimulation (EMS) by phase angle electrical impedance tomography,” Physiological Measurement, vol. 45, no. 7, pp. 1-15, Aug, 2024.
[30] P. N. Darma and M. Takei, “High-Speed and Accurate Meat Composition Imaging by Mechanically-Flexible Electrical Impedance Tomography With k-Nearest Neighbor and Fuzzy k-Means Machine Learning Approaches,” IEEE Access, vol. 9, pp. 38792–38801, Mar, 2021.
[31] T. Rymarczyk, G. Kłosowski, E. Kozłowski and P. Tchórzewski, “Comparison of Selected Machine Learning Algorithms for Industrial Electrical Tomography,” Sensors, vol. 19, no. 7, pp. 1521, Jan, 2019.
[32] F. A. Escobar Revelo, V. H. Mosquera Leyton and C. F. Rengifo Roda, “Electrical Impedance Tomography: Hardware Fundamentals And Medical Applications,” Ingenieria Solidaria, vol. 16, no. 3, pp. 1-29, Sep, 2020.
[33] R. Harikumar, R. Prabu and S. Raghavan, “Electrical Impedance Tomography (EIT) and Its Medical Applications : A Review,” International Journal of Soft Computing and Engineering, vol. 3, no. 4, pp. 193–198, Sep, 2013.
[34] T. K. Bera and J. Nagaraju, “A LabVIEW Based Data Acquisition System for Electrical Impedance Tomography (EIT),” Proceedings of the Third International Conference on Soft Computing for Problem Solving, vol. 259, pp. 377–389, Jan, 2014.
[35] T. K. Bera, “Wireless Electrical Impedance Tomography: LabVIEW-Based Automatic Electrode Switching,” Telehealth and Mobile Health, pp.639-665, Jan, 2016.
[36] A. Adler and W. R. B. Lionheart, “Uses and abuses of EIDORS: an extensible software base for EIT,” Physiological Measurement, vol. 27, no. 5, pp. 25–42, May, 2006.
[37] M. Vauhkonen, W. R. B. Lionheart, L. M. Heikkinen, P. J. Vauhkonen, and J. P. Kaipio, “A MATLAB package for the EIDORS project to reconstruct two-dimensional EIT images,” Physiological Measurement, vol. 22, no. 1, pp. 107–111, Feb, 2001.
[38] E. Ma, “An equilibrium finite element method for electrical impedance tomography,” Eng. Res. Express, vol. 5, no. 1, p. 1-11, Mar, 2023.
[39] T. K. Bera and J. Nagaraju, “A FEM-Based Forward Solver for Studying the Forward Problem of Electrical Impedance Tomography (EIT) with A Practical Biological Phantom” , IEEE International Advance Computing Conference, pp. 6–7, Mar, 2009.
[40] A. Lechleiter and A. Rieder, “Newton regularizations for impedance tomography: a numerical study,” Inverse Problems, vol. 22, no. 6, pp. 1967–1987, Dec, 2006.
[41] M. Vauhkonen, D. Vadasz, P. A. Karjalainen, E. Somersalo, and J. P. Kaipio, “Tikhonov regularization and prior information in electrical impedance tomography,” IEEE Transactions on Medical Imaging, vol. 17, no. 2, pp. 285–293, Apr, 1998.
[42] T. Chen and C. Guestrin, "Xgboost: A scalable tree boosting system," Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining, pp. 785-794, Mar, 2016.