The CAN is developed from an asymmetric convolutional neural network (CNN) with a complex-valued transformer encoder, harnessing the strengths of local feature extraction provided by the CNN and the ...

Second, the applied deep learning method is based on an autoencoder where a combination of Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) is utilized as the autoencoder ...