Real-Time people Identification and tracking based on deep learning

Authors

  • Muthana H. Hamd, Ansam H. Rashed

Abstract

A face is specified as a primitive center for consideration in our social lives since it is vital in
authentication and identification. Because of its non intrusiveness, providing fast and accurate results, real
time face recognition is of high importance in the systems of security. There are 2 levels for executing
real-time human face recognition, like face recognition and face detection. Because of its high real-time
permit rate and high-precision, the viola and jones algorithm is used for face detection, while there are 2
categories indicated in face recognition: the evaluating phase and training phase. In the latter, the
algorithm was trained with image samples which will be learned, whereas, in the former, a test image has
been put to comparison with all data set trained samples. In addition, the facial features have been
extracted with detected faces from the live stream via Principal Component Analysis (PCA), Linear
Discriminant Analysis (LDA), and Local Binary Patterns Histogram (LBPH). Face recognition has been
done via utilizing Convolutional Neural Networks Classifier (CNN). The Experimental results getting
after implementations and testing is of accuracy 100%.

Published

2020-12-02

Issue

Section

Articles