Comparison of iteratively regularized Gauss-Newton method with Adam optimization for image reconstruction in electrical impedance tomography

Abstract

Electrical impedance tomography (EIT) is a non-invasive imaging technique that visualizes the distribution of electrical conductivity within biological tissues. In this paper, we explore the potential of the Adam optimization algorithm as an innovative approach to reconstructing thoracic conductivity from EIT data. Originally developed for machine learning and signal processing tasks, the Adam method offers significant promise for enhancing EIT reconstruction accuracy and spatial resolution. Through comprehensive numerical simulations conducted on thorax models, we compare the Adam method and the traditional iterative Gauss-Newton method. The results demonstrate that the Adam method provides superior performance, improving spatial resolution and accuracy in resolving thoracic conductivity. The method is still under investigation, and further research and validation are needed to fully establish its effectiveness and reliability. Although preliminary findings are promising, additional research and clinical trials are required to identify the degree of its benefits and limits in the context of thoracic imaging. This study contributes to the growing body of research aimed at exploring advanced optimization methods to optimize EIT applications in medical imaging. This will result in better diagnostic capabilities and medical decision making in the field of thoracic health.