Performance Evaluation and Comparison of Software for Face Recognition, Based on Dlib and Opencv Library

Overview and investigate time complexity of computer vision algorithms for face recognition. Main article idea is to compare two popular computer vision librarieobjs, they are OpenCV and dlib, explore features, analyze pros and cons each of them and understand in what situation each of them suit the best. Method. The technologies of computer vision, which are used for face recognition was worked out. Research of two popular computer vision libraries was conducted. Their features are analyzed and the advantages and disadvantages of each of them are estimated. Examples of building recognition application based on histogram-oriented gradients for face finding, face landmark estimation for face orientation, and deep convolutional neural network to compare with known faces. The article generalizes the concept of face recognition. The scientific basis for facial recognition and the construction of a complete recognition system was described. The basic principles of the programs for face recognition are formulated. A comparative analysis of the productivity of both libraries in relation to - the time of execution to the number of iterations of the applied algorithms was presented. Also built two simple applications for face recognition based on these libraries and comparing their performance.