Development of a Thoracic Deformation Measurement Belt Based on Hall Sensors and a Deep Learning-Based Electrical Impedance Tomography Method for Lung Ventilation Considering Thoracic Deformation

Electrical impedance tomography (EIT), due to its non-radiative nature and high temporal resolution, holds significant clinical value in monitoring lung ventilation. However, this emerging imaging technology is still under development and has not been widely adopted. One of its major limitations is the low imaging resolution, which is susceptible to interference. Additionally, the accuracy of EIT imaging can be compromised by thoracic deformation during the monitoring of lung ventilation. To address these issues, a deep learning-based correction method utilizing a Hall sensor-equipped chest belt has been proposed for lung ventilation monitoring under conditions of thoracic movement. The chest belt, composed of Hall displacement sensors, converts displacement signals caused by respiratory movements into electrical signals. These signals, along with the corresponding thoracic impedance measurements, are fed into a dual-input deep convolutional network with a residual structure, producing corrected EIT images of lung ventilation. In simulation experiments, the proposed method significantly reduces imaging artifacts caused by lung ventilation movements and effectively corrects errors induced by thoracic deformation in EIT imaging.