Deep Learning CNN Convolutional Neural Networks with Python - Convolution and Pooling Details

The video tutorial explains a CNN example using a 32x32 grayscale image with two 5x5 kernels. It covers the convolution process, including applying Relu nonlinearity and zero padding, and details the computation of convolution entries. The tutorial also explains max pooling with a 2x2 window and stride, and concludes with a basic Python implementation of convolution and pooling using nested loops.

What is the size of the image used in the CNN example?

• 128x128

• 16x16

• 32x32

• 64x64

Which nonlinearity is applied after the convolution operation?

• Sigmoid

• Tanh

• ReLU

• Softmax

What is the purpose of zero padding in the convolution process?

• To change the color of the image

• To maintain the size of the image

• To increase the size of the image

• To decrease the size of the image

How is the dot product calculated in the convolution process?

• By subtracting kernel values from image values

• By adding all pixel values

• By dividing image values by kernel values

• By multiplying kernel and image entries and summing them

What is the main operation performed during max pooling?

• Summing all values

• Finding the maximum value

• Finding the minimum value

• Calculating the average value

In the Python implementation, which method is NOT used for convolution?

• Manual calculation

• Built-in convolve function

• Vectorized code

• Nested loops

What is the stride used in the max pooling process?

• 4

• 1

• 2

• 3