Method and device for generating a vessel imaging sequence
A vessel imaging sequence including plurality of vessel imaging frames and a corresponding ECG signal are generated by encoding, using a first encoder, an input vessel imaging sequence to generate a plurality of vessel latent space vectors, each vessel latent space vector corresponding to an input vessel imaging frame of the input vessel imaging sequence, encoding, using a second encoder, an input ECG signal to generate a plurality of ECG latent space vectors, decoding, using a first decoder, the plurality of vessel latent space vectors and the plurality of ECG latent space vectors to generate the vessel imaging sequence and decoding, using a second decoder, the vessel latent space vector and the ECG latent space vector to generate the ECG signal.
1 . A method for generating a vessel imaging sequence comprising a plurality of vessel imaging frames and a corresponding electrocardiogram (ECG) signal, the method comprising:
obtaining an input vessel imaging sequence comprising a plurality of input vessel imaging frames and a corresponding input ECG signal;
encoding, using a first encoder, the input vessel imaging sequence, the encoding using the first encoder generating a plurality of vessel latent space vectors, each vessel latent space vector corresponding to an input vessel imaging frame of the input vessel imaging sequence;
encoding, using a second encoder, the input ECG signal, the encoding using the second encoder generating a plurality of ECG latent space vectors;
decoding, using a first decoder, the plurality of vessel latent space vectors and the plurality of ECG latent space vectors, the decoding using the first decoder generating the vessel imaging sequence; and
decoding, using a second decoder, the vessel latent space vector and the ECG latent space vector, the decoding using the second decoder generating the ECG signal.
2 . The method of claim 1 , wherein a frame rate of the input vessel imaging sequence is lower than a frame rate of the vessel imaging sequence.
3 . The method of claim 2 , wherein the input vessel imaging sequence is obtained during a single cardiac cycle.
4 . The method of claim 3 , wherein the first encoder, the second encoder, the first decoder and the second decoder are each a convolutional neural network, a principal component analysis or a transformer-based neural network architecture.
5 . The method of claim 4 , wherein:
the first encoder and the first decoder form a first modified latent space autoencoder; and
the second encoder and the second decoder form a second modified latent space autoencoder.
6 . The method of claim 1 , wherein the input vessel imaging sequence is an input vessel image sequence, an input vessel segmentation sequence derived from the input vessel image sequence, or an input fluoroscopic imaging sequence.
7 . The method of claim 1 , wherein the input vessel imaging sequence is obtained during a single cardiac cycle.
8 . The method of claim 1 , wherein the obtaining the input vessel imaging sequence and the corresponding input ECG signal includes deriving the input ECG signal from the input vessel imaging sequence.
9 . The method of claim 1 , further comprising deriving at least one of a respiratory motion model and a cardiac motion model from the plurality of vessel latent space vectors and the plurality of ECG latent space vectors.
10 . The method of claim 1 , wherein the plurality of vessel latent space vectors and the plurality of ECG latent space vectors are respectively configured to enable a reconstruction of the input vessel imaging sequence and of the input ECG signal.
11 . The method of claim 1 , wherein the first encoder, the second encoder, the first decoder and the second decoder are each a convolutional neural network, a principal component analysis, or a transformer-based neural network architecture.
12 . The method of claim 1 , wherein:
the first encoder and the first decoder form a first modified latent space autoencoder; and
the second encoder and the second decoder form a second modified latent space autoencoder.
13 . A vessel imaging sequence generation system, the vessel imaging sequence including a plurality of vessel imaging frames, the vessel imaging sequence generation system comprising:
at least one processor; and
a storage medium comprising machine-readable instructions,
wherein the machine-readable instructions cause the at least one processor to:
obtain an input vessel imaging sequence comprising a plurality of input vessel imaging frames and a corresponding input ECG signal;
encode, using a first encoder, the input vessel imaging sequence to generate a plurality of vessel latent space vectors, each vessel latent space vector corresponding to an input vessel imaging of the input vessel imaging sequence;
encode, using a second encoder, the ECG signal to generate a plurality of ECG latent space vectors;
decode, using a first decoder, the plurality of vessel latent space vectors and the plurality of ECG latent space vectors to generate the vessel imaging sequence; and
decode, using a second decoder, the vessel latent space vector and the ECG latent space vector to generate the ECG signal.
14 . The vessel imaging sequence generation system of claim 13 , wherein a frame rate of the input vessel imaging sequence is lower than a frame rate of the vessel imaging sequence.
15 . The vessel imaging sequence generation system of claim 14 , wherein the input vessel imaging sequence is obtained during a single cardiac cycle.
16 . The vessel imaging sequence generation system of claim 15 , wherein the machine-readable instructions further cause the at least one processor to derive at least one of a respiratory motion model and a cardiac motion model from the plurality of vessel latent space vectors and the plurality of ECG latent space vectors.
17 . The vessel imaging sequence generation system of claim 13 , wherein the input vessel imaging sequence is an angiographic imaging sequence, a vessel segmentation sequence derived from the angiographic imaging sequence, or a fluoroscopic imaging sequence.
18 . The vessel imaging sequence generation system of claim 13 , wherein the input vessel imaging sequence is obtained during a single cardiac cycle.
19 . The vessel imaging sequence generation system of claim 13 , wherein, to obtain the input vessel imaging sequence and the corresponding input ECG signal, the machine-readable instructions further cause the at least one processor to derive the input ECG signal from the input vessel imaging sequence.
20 . The vessel imaging sequence generation system of claim 13 , wherein the machine-readable instructions further cause the at least one processor to derive at least one of a respiratory motion model and a cardiac motion model from the plurality of vessel latent space vectors and the plurality of ECG latent space vectors.