Conditional Diffusion Model for Electrical Impedance Tomography

TL;DR

This work tackles the ill-posed, noise-sensitive inverse problem in electrical impedance tomography by proposing a conditional diffusion framework (CDMVC) that uses an initial PDIPM-generated reconstruction as a conditioning input for a VP-SDE diffusion process. A forward voltage constraint network enforces alignment between the diffusion-generated conductivity and boundary voltages during sampling, improving image fidelity and realism. The approach is supported by a large EIT dataset built with EIDORS, extensive simulations, and physical tactile-sensor experiments, demonstrating strong anti-noise performance and generalization to new shapes. Results show CDMVC outperforms PDIPM, CNN/GAN-based reconstructions, and prior diffusion-based post-processing, offering a robust, high-quality reconstruction pipeline with practical potential for real-time or near-real-time EIT imaging tasks.

Abstract

Electrical impedance tomography (EIT) is a non-invasive imaging technique, which has been widely used in the fields of industrial inspection, medical monitoring and tactile sensing. However, due to the inherent non-linearity and ill-conditioned nature of the EIT inverse problem, the reconstructed image is highly sensitive to the measured data, and random noise artifacts often appear in the reconstructed image, which greatly limits the application of EIT. To address this issue, a conditional diffusion model with voltage consistency (CDMVC) is proposed in this study. The method consists of a pre-imaging module, a conditional diffusion model for reconstruction, a forward voltage constraint network and a scheme of voltage consistency constraint during sampling process. The pre-imaging module is employed to generate the initial reconstruction. This serves as a condition for training the conditional diffusion model. Finally, based on the forward voltage constraint network, a voltage consistency constraint is implemented in the sampling phase to incorporate forward information of EIT, thereby enhancing imaging quality. A more complete dataset, including both common and complex concave shapes, is generated. The proposed method is validated using both simulation and physical experiments. Experimental results demonstrate that our method can significantly improves the quality of reconstructed images. In addition, experimental results also demonstrate that our method has good robustness and generalization performance.

Figures (11)

• Figure 1: It includes four parts. As for inference phase, the input consists of the initially reconstructed conductivity concatenated with a noise sample from Gaussian noise. This concatenated feature is fed into the DDIM sampling process, where voltage consistency constraint optimization is applied to outputs of the intermediate steps.
• Figure 2: Voltage measurements under different settings
• Figure 3: The process of generating training data
• Figure 4: Visual results of different algorithms
• Figure 5: The ablation experiment of our method