Rozpoznávání tváří pomocí konvolučních neuronových sítí

Abstract

This master thesis deals with face recognition using convolutional neural networks. In the theoretical part, convolutional networks are described together with methods and architectures, which allow training of deep and powerful neural network models. All these methods and architectures are used for training many models, which are then tested and compared on benchmark dataset for face verification called Labeled Faces in the Wild.