Tutorial by Topics: convolution

There are many ways of visualizing the convolutional layers, but they share same components: fetching the values of a part of the convolutional neural networks, and visualizing those values. Note those visualizations should not and can not display on the TensorBoard.

2D convolution is computed in a similar way one would calculate 1D convolution: you slide your kernel over the input, calculate the element-wise multiplications and sum them up. But instead of your kernel/input being an array, here they are matrices.