The Dropout Method of Face Recognition Using a Deep Convolution Neural Network
Dian Yi, Shi Xiaohong, Hao Xu · 2018
With the rapid development of artificial intelligence and pattern recognition, face recognition has become a hot topic in the field of computer vision. Especially after the deep learning proposed, the performance of face recognition algorithm has been greatly improved. This paper mainly introduces the main method of face recognition using a deep convolution neural network model. First, for training network parameters, we use a fast convergence stochastic gradient algorithm (SGD). At the same time, the "dropout" method is added to each layer of the network to hide some neuron activity by a certain probability, in order to avoid the over fitting problem caused by the deep network model. Through the above process, a neural network model can recognize face images.