VDD-Net-based lung electrical impedance imaging method

Abstract

The invention provides a lung electrical impedance imaging method based on VDD-Net. The lung electrical impedance imaging method comprises a conjugate gradient pre-mapping module, a feature extraction module, a depth image reconstruction module and an image denoising module. The depth image reconstruction module maps the measured voltage sequence into field domain spatial distribution information, so that an EIT ill-conditioned problem is converted into a good problem, abstract features of spatial information are extracted through multilayer convolution, and a lung boundary shape is reconstructed by using transposed convolution operation, so that the lung boundary shape is reconstructed, and the reconstructed image is connected with a low-pass filtering module through Dense to remove reconstructed high-frequency noise, and a final reconstructed image is obtained. The training data of VDD-Net uses CT scanning images containing various lung boundary information, and noise conditions such as lung diseases, different thoracic cavity shapes and electrode movement are included. A lung phantom model experiment shows that the VDD-Net has relatively high accuracy on boundary reconstruction capability in lung EIT imaging and has good robustness on model errors and measurement noise.

Description

technical field [0001] The invention belongs to the field of electrical tomography. On the basis of the convolutional neural network, the four sub-modules of pre-reconstruction, feature extraction, image reconstruction and image denoising are connected in sequence, which are mainly used in tomography of lung respiratory status, auxiliary diagnosis of some lung diseases and visualization of lung image reconstruction field. Background technique [0002] Lung imaging is a relatively hot area of ​​research in today's imaging field. The large-scale outbreak of the new crown epidemic has indirectly promoted scientists around the world to actively explore new imaging methods. Traditional imaging equipment and imaging methods lack strong real-time performance. When patients need further condition monitoring after being infected with the new coronavirus, the lack of real-time imaging methods often leads to delays in their condition. Currently, the main imaging techniques on the mar...

AI Technical Summary

Problems solved by technology

However, the relevant parameters of iterative algorithms are mostly difficult to choose. The number of iterations needs to be determined by experience and human experience. The sensitivity matrix is ​​also very dependent on the accuracy of the positive problem model and prior information, which greatly limits the scope of EIT imaging technology. Generalization ability also limits its large-scale clinical application

Images