Table 3 Modified parameters in the convolutional auto-encoder network for the classification of four categories
Layer (type) | Output shape | Param # |
|---|---|---|
Input_1 (InputLayer) | (None, 80, 80, 3) | 0 |
Conv2d (Conv2D) | (None, 80, 80, 128) | 1664 |
Batch_normalization (BatchNormalization) | (None, 80, 80, 128) | 512 |
Conv2d_1 (Conv2D) | (None, 80, 80, 128) | 65,664 |
Batch_normalization_1 (BatchNormalization) | (None, 80, 80, 128) | 512 |
Max_pooling2d (MaxPooling2D) | (None, 40, 40, 128) | 0 |
Dropout (Dropout) | (None, 40, 40, 128) | 0 |
Conv2d_2 (Conv2D) | (None, 40, 40, 64) | 32,832 |
Batch_normalization_2 (BatchNormalization) | (None, 40, 40, 64) | 256 |
Conv2d_3 (Conv2D) | (None, 40, 40, 64) | 16,448 |
Batch_normalization_3 (BatchNormalization) | (None, 40, 40, 64) | 256 |
Max_pooling2d_1 (MaxPooling2D) | (None, 20, 20, 64) | 0 |
Dropout_1 (Dropout) | (None, 20, 20, 64) | 0 |
Conv2d_4 (Conv2D) | (None, 20, 20, 32) | 8224 |
Batch_normalization_4 (BatchNormalization) | (None, 20, 20, 32) | 128 |
Max_pooling2d_2 (MaxPooling2D) | (None, 10, 10, 32) | 0 |
Conv8 (Conv2D) | (None, 10, 10, 64) | 8256 |
Conv2d_11 (Conv2D) | (None, 10, 10, 64) | 16,448 |
Batch_normalization_10 (BatchNormalization) | (None, 10, 10, 64) | 256 |
Max_pooling2d_7 (MaxPooling2D) | (None, 5, 5, 64) | 0 |
Dropout_5 (Dropout) | (None, 5, 5, 64) | 0 |
Flatten (Flatten) | (None, 1600) | 0 |
Dense (Dense) | (None, 4) | 6404 |
Total parameters: 157, 860 | ||
Trainable parameters: 156, 900 | ||
Non-trainable parameters: 960 | ||