I'm working on a project about graph neural networks and was reading some stuff about convolutional neural networks. I did not understand what the book means with 'preserve local connectivity', because in the next sentence they say that it contains fully connected layers as well.
Convolutional neural networks (CNNs) are special versions of feedworward neural networks (FNNs). FNNs are usually fully connected networks while CNNs preserve the local connectivity. The CNN architecture usually contains convolutional layers, pooling layers, and several fully connected layers
I was hoping someone could explain this to me :)
Edit : I found a nice answer here : https://stats.stackexchange.com/questions/159588/how-does-local-connection-implied-in-the-cnn-algorithm#:~:text=Each%20input%20(pixel%20value)%20is,EVERY%20part%20of%20the%20image.&text=This%20is%20what%20is%20meant,actually%20close%20to%20each%20other.