Question

Just want to see if I get these True or False answered correctly:

True - Since we apply the same convolutional filter throughout the image, we can learn to recognize the same feature wherever it appears.

False - A fully connected layer for a reasonably sized image would simply have too many parameters

False - A fully connected layer can learn to recognize features anywhere in the image even if the features appeared preferentially in one location during training

Answer 1

Since we apply the same convolution filter throughout the image, we can learn to recognize the same feature wherever it appears: This statement is true as the convolution matrix is used to detect the same features in an image by the help of the feature map.

A fully connected layer for a reasonably sized image would simply have too many parameters: This statement is true as in the fully connected layer each neuron is connected to every other neuron. So, the number of parameters is the product of number of neurons in current layer and the neuron number of the previous layer plus one. So, for a image of big or medium size there will be too many parameters.

A fully connected layer can learn to recognize features anywhere in the image even if the features appeared preferentially in one location during training: This statement is true. If the fully connected layer is trained with the feature then it can recognize it. In fully connected layer all the neurons are interconnected. This characteristics helps the fully connected layer to recognize features anywhere in the image even if the features appeared preferentially in one location during training.

Hope this helps.