Design of Face Recognition System Based on Convolutional Neural Network

Abstract

This paper designs a service robot-oriented face recognition system. For mobile robots, face recognition is a very important function among them. The system includes three parts: face acquisition and preprocessing, model establishment, and model training. Among them, the face collection link uses the face detection function in opencv, and optimizes the interference factors to establish its own data set. A convolutional neural network was designed and constructed, and the model was trained on its own data set. The accuracy rate obtained on the test set was 97.63%. Finally, the trained model was applied to the actual system. The system model is simple, occupies a small amount of memory, and can be applied to the actual application scenario of the robot moving forward, which can quickly and accurately detect and recognize human faces.