Physics-Driven Neural Compensation For Electrical Impedance Tomography

2025-04-25

AI Summary

EIT reconstruction faces two fundamental challenges: ill-posedness of the inverse problem and spatially non-uniform sensitivity. Conventional model-based methods rely on handcrafted regularization but neglect sensitivity heterogeneity, while supervised learning suffers from poor generalizability. Existing neural field approaches lack explicit physical constraints, limiting reconstruction accuracy. To address these issues, we propose an unsupervised, physics-driven neural compensation framework. Our method introduces a novel dynamic neural representation allocation mechanism that adaptively enhances representational capacity in low-sensitivity regions, guided by the sensitivity distribution derived from the EIT forward model. It integrates the physical forward model for implicit regularization and incorporates sensitivity-aware coordinate encoding with adaptive weight modulation. Evaluated on both synthetic and experimental data, our approach significantly improves structural fidelity and artifact suppression—particularly mitigating distortion in low-sensitivity regions—and demonstrates superior robustness compared to state-of-the-art model-based and supervised learning methods.

Technology Category

Computer Vision: Low Level & Physics-based VisionMachine Learning: Calibration & Uncertainty QuantificationCognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystemsGraph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphsUser Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Electrical Impedance Tomography (EIT) provides a non-invasive, portable imaging modality with significant potential in medical and industrial applications. Despite its advantages, EIT encounters two primary challenges: the ill-posed nature of its inverse problem and the spatially variable, location-dependent sensitivity distribution. Traditional model-based methods mitigate ill-posedness through regularization but overlook sensitivity variability, while supervised deep learning approaches require extensive training data and lack generalization. Recent developments in neural fields have introduced implicit regularization techniques for image reconstruction, but these methods typically neglect the physical principles underlying EIT, thus limiting their effectiveness. In this study, we propose PhyNC (Physics-driven Neural Compensation), an unsupervised deep learning framework that incorporates the physical principles of EIT. PhyNC addresses both the ill-posed inverse problem and the sensitivity distribution by dynamically allocating neural representational capacity to regions with lower sensitivity, ensuring accurate and balanced conductivity reconstructions. Extensive evaluations on both simulated and experimental data demonstrate that PhyNC outperforms existing methods in terms of detail preservation and artifact resistance, particularly in low-sensitivity regions. Our approach enhances the robustness of EIT reconstructions and provides a flexible framework that can be adapted to other imaging modalities with similar challenges.

Problem

• Addresses ill-posed inverse problem in EIT imaging
• Compensates for spatially variable sensitivity distribution
• Enhances robustness without requiring extensive training data

Innovation

• Unsupervised deep learning for EIT reconstruction
• Dynamic neural capacity allocation for sensitivity
• Physics-driven neural compensation (PhyNC) framework