From: Semi-supervised underwater acoustic source localization based on residual convolutional autoencoder
Block | Output channel | Kernel size | Stride | |
---|---|---|---|---|
Conv1 | 64 | \({3} \times {3}\) | 1 | |
Encoder | Conv2 | 128 | \({3} \times {3}\) | 2 |
Conv3 | 256 | \({3} \times {3}\) | 2 | |
Transposed-Conv1 | 256 | \({3} \times {3}\) | 2 | |
Decoder | Transposed-Conv2 | 128 | \({2} \times {2}\) | 2 |
Transposed-Conv3 | 64 | \({4} \times {4}\) | 1 | |
Dense_1 | 4096 | - | 1 | |
MLP | Dense_2 | 2048 | - | 1 |
Dense_3 | 1024 | - | 1 | |
Dense_4 | 100 | - | 1 |