Table 1 LCA Net architecture.
From: ODD-Net: a hybrid deep learning architecture for image dehazing
Layer | Output shape | Trainable parameters |
|---|---|---|
Input_layer | (Batch, 352, 1216, 3) | 0 |
Conv2d_1(encoder) | (Batch, 352, 1216, 50) | 1400 |
Average_pool_1 (encoder) | (Batch, 176, 608, 50) | 0 |
Conv2d_2(encoder) | (Batch, 176, 608, 50) | 22,550 |
Average_pool_2 (encoder) | (Batch, 88, 304, 50) | 0 |
Dense_1 (encoder) | (Batch, 88, 304, 10) | 510 |
Dense_2 (encoder) | (Batch, 88, 304, 10) | 110 |
Conv2d_Transpose_1 (decoder) | (Batch, 88, 304, 50) | 4550 |
UpSample2d_1 (decoder) | (Batch, 176, 608, 50) | 0 |
Conv2d_Transpose_2 (decoder) | (Batch, 176, 608, 50) | 22,550 |
UpSample2d_2 (decoder) | (Batch, 352, 1216, 50) | 0 |
Conv2d_Transpose_3 (decoder) | (Batch, 352, 1216, 3) | 1353 |
