Design And Implementation Of Face Recognition System Based On Robot Scene Interaction

For now,it is a new era of intelligent robots explosion.Many kinds of robots,such as intelligent robots,servicing robots,recreational robots and educational robots always bring surprise to people and society.In this high speed rhythm life,parents always busy with their works and have no time to accompany with their children.Intelligent children education robots help parents company with children,and provide better education for children.Face recognition system based on robot scene interaction is an important module of intelligent children education robot.The purpose is to help robots identify different faces and have better interaction with people.Face recognition system based on robot scene interaction is applied in the specific scene of people's life.As we all know that life scenes are rich and varied,so we need to enrich its scene and improve processing capacity as much as possible.This problem needs to be resolved from two aspects,on the one hand can increase the scene logic as much as possible;on the other hand,to improve the applicability of the scene,with a small amount of scene logic to deal with more scene.This paper uses face recognition technology to realize two scenes:one is meeting new friends and the other is saying "hello" to good friends.Firstly,this paper introduces practical application background of this project and how face recognition used in this system.At the same time,this paper also introduced the present situation of face recognition at home and abroad.Secondly,combined with the business needs of the system to determine the overall system architecture and module division,as well as the interaction between modules.Finally,this paper elaborates the realization process of face recognition system based on robot scene interaction,which is the most important part of this paper.This innovation of this paper includes the following aspects:firstly,this article applied face recognition algorithm in real interactive scenario;secondly,in the feature extraction algorithm,use the output of CNN's full connection,this method not only retained the CNN identification accuracy,and improve the efficiency of feature extraction;finally,the separation of the scene and visual algorithm,improved the system scalability.