Table 6:

Description of our CNN to the MNIST dataset

Layer	Description
Convolution	Input image 28 × 28, kernel size 7 × 7, stride size of 3, number of output channels 4
1st square	Squaring 256 input values
FC-1	Fully connecting with 256 inputs and 64 outputs: ${ℝ}^{64\times 256}\times {ℝ}^{256\times 1}\to {ℝ}^{64\times 1}$
2nd square	Squaring 64 input values
FC-2	Fully connecting with 64 inputs and 10 outputs: ${ℝ}^{10\times 64}\times {ℝ}^{64\times 1}\to {ℝ}^{10\times 1}$