Have you ever uploaded a photo of your friends' group to Facebook? If yes, have you ever wondered how Facebook detects all the faces in the photo automatically just after the upload finishes? In short, the answer is convolutional neural networks (CNNs).
A feed-forward network generally consists of several fully connected layers, whereas a CNN consists of several layers of convolution, along with other types of sophisticated layers, including fully-connected layers. These fully-connected layers are generally placed towards the very end and are typically used for making predictions. But what kinds of predictions? In an image-processing and computer-vision context, a prediction task can encompass many use cases, ...
