Below is the summary of our CNN model which we have made using python and keras.
| Layer (type) | Output Shape | Param # |
|---|---|---|
| conv2d (Conv2D) | (None, 222, 222, 32) | 896 |
| max_pooling2d (MaxPooling2D) | (None, 111, 111, 32) | 0 |
| conv2d_1 (Conv2D) | (None, 109, 109, 64) | 18496 |
| max_pooling2d_1 (MaxPooling2D) | (None, 54, 54, 64) | 0 |
| conv2d_2 (Conv2D) | (None, 52, 52, 128) | 73856 |
| max_pooling2d_2 (MaxPooling2D) | (None, 26, 26, 128) | 0 |
| flatten (Flatten) | (None, 86528) | 0 |
| dense (Dense) | (None, 128) | 11075712 |
As you can see this is a simple Neural Network with 8 layers. (this had an accuracy of approx 0.93 on average) Each Layer has a specific job in order to get the desired output, they are :-
-
Convolutional layer (conv2d): This layer takes an input image and applies a set of 32 filters to produce 32 output feature maps. Each filter extracts a particular feature from the image. The output of this layer has a shape of (None, 222, 222, 32).
-
Max pooling layer (max_pooling2d): This layer reduces the dimensionality of the output of the previous layer by taking the maximum value in each 2x2 region. The output of this layer has a shape of (None, 111, 111, 32).
-
Convolutional layer (conv2d_1): This layer applies a set of 64 filters to the output of the previous layer to produce 64 output feature maps. The output of this layer has a shape of (None, 109, 109, 64).
-
Max pooling layer (max_pooling2d_1): This layer reduces the dimensionality of the output of the previous layer by taking the maximum value in each 2x2 region. The output of this layer has a shape of (None, 54, 54, 64).
-
Convolutional layer (conv2d_2): This layer applies a set of 128 filters to the output of the previous layer to produce 128 output feature maps. The output of this layer has a shape of (None, 52, 52, 128).
-
Max pooling layer (max_pooling2d_2): This layer reduces the dimensionality of the output of the previous layer by taking the maximum value in each 2x2 region. The output of this layer has a shape of (None, 26, 26, 128).
-
Flatten layer (flatten): This layer flattens the output of the previous layer into a 1D vector. The output of this layer has a shape of (None, 86528).
-
Fully connected layer (dense): This layer takes the flattened vector from the previous layer and applies 128 neurons to it, producing a 128-dimensional output. The output of this layer has a shape of (None, 128).