Imaging device of eliminating electromagnetic interference of magnetic resonance and imaging method thereof

An imaging method of eliminating electromagnetic interference of magnetic resonance, applied to an imaging device of eliminating electromagnetic interference of magnetic resonance, comprising:

• Step 1, the imaging device acquiring a plurality of magnetic resonance imaging signals, acquiring a first electromagnetic interference signal, by a receiving coil, obtained by measurement in an electromagnetic interference affected environment, and acquiring second electromagnetic interference signals, by a plurality of sensing coils, obtained by measurement in the electromagnetic interference affected environment;
• Step 2, the imaging device superposing each of the plurality of magnetic resonance imaging signals with the first electromagnetic interference signal to obtain interfered imaging signals, respectively, taking the interfered imaging signals and corresponding second electromagnetic interference signals as input, and taking the plurality of magnetic resonance imaging signals as output, to train an electromagnetic interference eliminating model;
• Step 3, the imaging device inputting a real-time magnetic resonance imaging signal that is acquired by the receiving coil and a real-time electromagnetic interference signal that is acquired by the plurality of sensing coils into the electromagnetic interference eliminating model in a magnetic resonance imaging process, to obtain a predicted magnetic resonance imaging signal that eliminates electromagnetic interference;
• and
• Step 4, the imaging device performing image reconstruction on the predicted magnetic resonance imaging signal to obtain a magnetic resonance image;
• a model structure of the electromagnetic interference eliminating model comprises a first convolution layer, a first residual network layer, a first strided convolution layer, a second residual network layer, a second strided convolutional layer, a third residual network layer, a third strided convolution layer, a fourth residual network layer, a first deconvolution layer, a fifth residual network layer, a second deconvolution layer, a sixth residual network layer, a third deconvolution layer, a seventh residual network layer, and a second convolution layer,
• an input of the first convolution layer is used as an input of the electromagnetic interference eliminating model;
• an input of the first residual network layer is connected to an output of the first convolution layer;
• an input of the first strided convolution layer is connected to an output of the first residual network layer;
• an input of the second residual network layer is connected to an output of the first strided convolution layer;
• an input of the second strided convolution layer is connected to an output of the second residual network layer;
• an input of the third residual network layer is connected to an output of the second strided convolution layer;
• an input of the third strided convolution layer is connected to an output of the third residual network layer;
• an input of the fourth residual network layer is connected to an output of the third strided convolution layer;
• an input of the first deconvolution layer is connected to an output of the fourth residual network layer;
• an input of the fifth residual network layer is connected to an output of the first deconvolution layer;
• an input of the second deconvolution layer is connected to an output of the fifth residual network layer;
• an input of the sixth residual network layer is connected to an output of the second deconvolution layer;
• an input of the third deconvolution layer is connected to an output of the sixth residual network layer;
• an input of the seventh residual network layer is connected to an output of the third deconvolution layer;
• an input of the second convolution layer is connected to an output of the seventh residual network, and an output of the second convolution layer is used as an output of the electromagnetic interference eliminating model;
• and
• the first strided convolution layer is in skip connection with the third strided convolution layer, the second strided convolution layer is in skip connection with the second strided convolution layer, and the third strided convolution layer is in skip connection with the first strided convolution layer.