Facial Recognition System Demo (CNN+Support Vector Machines+GPU)

Nowadays, Closed Circuit Television (CCTV) monitoring systems, access control, and many other security-related applications incorporate facial recognition techniques. This disruptive tool differs from other biometric techniques, since faces can be recognized remotely. Thus, these applications can be incorporated in different institutions with the purpose of restricting access to unauthorized/unknown people, avoiding damages and losses to the public and private goods. The objective of this work is to identify people in controlled and uncontrolled environments within a university building that has suffered insecurity problems on several occasions. For this, a convolutional neural network (CNN) architecture implemented on a GPU was used. The procedure consists of 3 phases: (1) training; (2) learning and; (3) classification. The CNN training stage was performed using the VGGFace2 dataset. The learning and generalization of deeply discriminatory facial embeddings of 512 bytes per face by joint supervision of softmax loss and center loss signals. The classification used a support vector machine (SVM) in several experiments with different amounts of classes. The efficiency of this approach in the mentioned environments in real time, using a sample of 128 students, was tested by means of quantitative metrics derived by confusion matrix. The overall accuracy in controlled and uncontrolled contexts was 96\% and 71.43\%, respectively.

[satyamrout]

Sir can u pls share the code it will be very much helpful