Register or Log In

International Society for Industrial Process Tomography

11th World Congress on Industrial Process Tomography

Image Enhancement for Electrical Impedance Tomography (EIT) Reconstruction from Piezoresistive Fabric using CycleGAN and pix2pixGAN

Felipe Alberto Solano Sanchez1, Anil Kumar Khambampati1, Kyung Youn Kim1*

1 Department of Electronic Engineering, Jeju National University, Jeju, South Korea

*Email: kyungyk@jejunu.ac.kr


ABSTRACT


Fetal health monitoring is crucial in prenatal care, and various techniques have been developed to assess fetal movements. However, these methods are often expensive and limited to clinical settings. This study explores the potential of using Electrical Impedance Tomography (EIT) with a piezoresistive fabric as a low-cost and non-invasive imaging method for fetal monitoring. The combination of EIT with wearable e-textile devices allows continuous and portable monitoring. To enhance the quality of EIT reconstructed images, this study proposes the use of CycleGAN and pix2pixGAN, which are deep learning models based on generative adversarial networks (GANs). The models learn a mapping between reconstructed images and target images, improving reconstruction accuracy and image quality. The results demonstrate the effectiveness of the proposed method in handling noisy data and achieving enhanced image generation. This research offers a promising approach to fetal monitoring using EIT and deep learning techniques, opening possibilities for affordable and accessible prenatal care.

Keywords: Electrical impedance tomography; image reconstruction; generative adversarial network, piezoresistive fabric

Industrial Application: General

Sign-in to access the full text