The invention discloses a three-dimensional pulmonary ventilation
electrical impedance imaging method and
system, and belongs to the technical field of
electrical impedance imagines.The method comprises the steps that a denoising
voltage measurement value is obtained through a residual denoising network; the random input quantity is sent to a dual-output neural network with an
encoder and a decoder, a full-connection
branch is added to the output end of the
encoder, a
latent variable is output, and a
lung mask is generated through a
lung mask generator and serves as model generation priori; the decoder generates a
lung ventilation image, and the dual-output
network structure provides implicit regularization; based on the denoising
voltage measurement value, the lung shape
mask and the estimated boundary
voltage, an image reconstruction
loss function containing a
data consistency item and a lung shape loss item is adopted, dual-output neural network parameters are optimized through minimum iteration of the reconstruction
loss function, and image reconstruction is conducted on each instance in an unsupervised mode. According to the method, the robust performance and generalization performance of
lung ventilation imaging are improved, and the
image quality is improved.