Fig. 2From: Semi-supervised underwater acoustic source localization based on residual convolutional autoencoderStructure of self-attention modelBack to article page