THE SOLUTION OF THE PURPOSE OF RECOGNITION OF PERSONS WITH THE USE OF MACHINE TRAINING ALGORITHMS

The purpose of this article is to generalize the solutions of experience and implement a neural network for face recognition. The neural network is based on special algorithms of machine learning. As an input, the algorithm receives an image with the face of one person or persons of several people, after which all persons in this image are searched using gradient histograms, the result is a fragment of the image where the basic structures of the face or persons are clearly seen. In order to determine the unique features of the face, it is necessary to take into account the difference in the angle of the face and the degree of its illumination, for this purpose, on selected fragments of the application of estimation algorithms to search for 68 points that exist on each face, it is possible to center the eyes and mouth as best as possible for more exact encoding. Encoding an image involves building an accurate "face map" consisting of 128 dimensions. Based on the search results, the convolutional neural network, using the SVM linear classifier algorithm, can determine the correspondence between different photos.

1. Histograms of Oriented Gradients for Human Detection (In CVPR'05). N. Dalal and B. Triggs. An effective pedestrian detector based on evaluating histograms of oriented image gradients in a grid. [Elektronnyy resurs] // URL: http://lear.inrialpes.fr/people/triggs/pubs/Dalal-cvpr05.pdf (data obrashcheniya: 25.12.2017).

2. V. Kazemi and S. Josephine. One millisecond face alignment with an ensemble of regression trees. In CVPR, 2014. [Elektronnyy resurs] // URL: http://www.csc.kth.se/~vahidk/papers/KazemiCVPR14.pdf (data obrashcheniya: 25.12.2017).

3. F. Schroff, D. Kalenichenko, and J. Philbin. Facenet: A unified embedding for face recognition and clustering. In Proc. CVPR, 2015. [Elektronnyy resurs] // URL: https://www.cvfoundation.org/openaccess/content_cvpr_2015/app/1A_089.pdf (data obrashcheniya: 25.12.2017).

4. Christopher M. Bishop F.R.Eng. Pattern Recognition and Machine Learning. [Elektronnyy resurs] // URL: https://goo.gl/WLqpHN (data obrashcheniya: 25.12.2017)