Feature learning algorithms find the common patterns that are important to distinguish between classes and extract them automatically to be used in a classification or regression process. Feature learning can be thought of as Feature Engineering done automatically by algorithms. In deep learning, convolutional layers are exceptionally good at finding good features in images to the next layer to form a hierarchy of nonlinear features that grow in complexity (e.g. blobs, edges -> noses, eyes, cheeks -> faces). The final layer(s) use all these generated features for classification or regression (the last layer in a convolutional net is, essentially, multinomial logistic regression).