Deep Learning is an algorithm based on some ideas from the brain to absorb many layers of expression, both concrete and abstract, thereby clarifying the meaning of data types. Deep Learning is applied in image recognition, speech recognition, natural language processing. Currently, many identification problems are solved by deep learning based methods thanks to its ability to solve problems in a large number of variables, large input size with superior performance and accuracy as compared to traditional classification methods, and its ability to build intelligent systems with high accuracy. In this article, the authors conducted a study into the convolutional neural network (CNN) which is one of the advanced deep learning models for the problem of facial recognition from video.