The convolution kernel is used to obtain the output feature map by sliding over the input image and multiplying it element by element with the image and then summing the results. By stacking multiple ...