A Method to Suppress Contact Impedance Noise Under a Framework of CT Image-Guided EIT for Pulmonary Ventilation Monitoring

The performance of an electrical impedance tomography (EIT) system in prolonged pulmonary ventilation monitoring is critically influenced by contact impedance noise (CIN). This article proposes a method for suppressing CIN within a framework of computed tomography (CT) image-guided EIT, aiming to enhance the imaging quality of pulmonary ventilation. First, a full-link noise model is introduced, linking the EIT system noises to image artifacts. The distribution characteristics of artifacts caused by different types of noise were then analyzed, revealing that CIN primarily induces edge-concentrated artifacts. Second, a CT image-guided EIT imaging scheme based on cosine similarity is developed to improve image quality and separate the edge artifacts. The inverse problem is solved using the Gauss–Newton iteration, which can be computationally reduced to the conjugate gradient (CG) method for real-time reconstruction. Finally, stemming from the full link noise model and CT-guided EIT (CT-EIT) scheme, an adaptive method is proposed to estimate and suppress the CIN in real-time from the artifact distribution in an EIT image. A series of simulations and experimental tests is carried out to validate the performance of the proposed method. It is indicated that the severe artifacts caused by the CIN can be significantly suppressed, with the accuracy of lung shape improved considerably. Overall, the proposed method can serve as an adaptive online approach to enhance and maintain the EIT imaging quality and stability of pulmonary ventilation monitoring, especially when long-term monitoring is required, where the contact noise is inevitable due to unconstrained patient movement.